The viscosity of nanofluids: a review of the theoretical, empirical and

numerical models

JOSUA P. MEYER, SAHEED A. ADIO, MOHSEN SHARIFPUR and PAUL N. NWOSU

Nanofluids Research Laboratory, Thermofluids Research Group, Department of Mechanical and

Aeronautical Engineering, University of Pretoria, Pretoria 0002, South Africa.

ABSTRACT

The enhanced thermal characteristics of nanofluids have made it one of the fastest-growing

research areas in the last decade. Numerous researches have shown the merits of nanofluids in

heat transfer equipment. However, one of the problems is the increase in viscosity due to the

suspension of nanoparticles. This viscosity increase is not desirable in the industry, especially

when it involves flow, such as in heat exchanger or micro-channel applications where lowering

pressure drop and pumping power are of significance. In this regard, a critical review of the

theoretical, empirical and numerical models for effective viscosity of nanofluids was presented.

Furthermore, different parameters affecting the viscosity of nanofluids such as nanoparticle

volume fraction, size, shape, temperature, pH and shearing rate were reviewed. Other properties

such as nanofluid stability and magnetorheological characteristics of some nanofluids were also

reviewed. The important parameters influencing viscosity of nanofluids are temperature,

nanoparticle volume fraction, size, shape, pH and shearing rate. Regarding the composite of

nanofluids, which can consist of different fluid bases and different nanoparticles, different accurate

correlations for different nanofluids need to be developed. Finally, there is a lack of investigation

into the stability of different nanofluids when the viscosity is the target point.

Address correspondence to Mohsen Sharifpur, Nanofluids Research Laboratory, Department of

Mechanical and Aeronautical Engineering, University of Pretoria, Pretoria, South Africa.
E-mail: mohsen.sharifpur@up.ac.za
Tel: +27 12 420 2448
Fax: +27 12 420 6632

1
INTRODUCTION

Colloidal suspension dates back to Maxwell’s study in 1873 [1]. Though the idea behind his study

was vivid, the imposed problems were too enormous for profitable engineering solutions [2]. In

1995, Choi [3] came up with a pioneering idea based on Maxwell’s study and suspended ultrafine

particles (nanoparticles) in conventional heat transfer fluids. His invention has opened up a myriad

of opportunities in research and development. Nanofluids descriptively are colloidal suspensions

containing metallic (Ag, Au, Al, Cu, Ni, etc.), non-metallic (single- and multi-wall carbon

nanotube (SWCNT and MWCNT), Si, Graphene, etc.), metallic oxide (Al2O3, CuO, NiO2,TiO2,

etc.) and oxides of non-metals (SiO2, SiC, MgO, CaCO3, etc.) nanoparticles suspended in

conventional heat transfer fluids such as water, engine oil, ethylene glycol, transformer oil, gear oil

or mixture of two or more heat transfer fluids [2, 3]. When compared with previous microparticle

suspensions in conventional heat transfer fluids, it is a special type of fluid with numerous

applications potentials because of its enhanced thermal conductivity, stability and homogeneity [3,

4]. Microscale particles in suspensions lead to abrasion, clogging of flow paths, pressure drop and

high pumping power requirements, therefore, its sustainability was impossible. Besides, nanofluids

can reduce the pumping power in engineering equipment significantly and do not pose the problem

of clogging and abrasion of equipment flow paths [5–8]. Therefore, the design and engineering of

physical systems are now being tailored towards using nanofluid as working fluid.

The impact of colloidal suspension cuts across the fields of science, biological science, medical,

pharmaceutical and engineering. In the context of sustainable energy development and thermal

management, nanofluids are becoming more and more significant as the need for efficient thermal

management is of paramount importance. Moreover, the level of miniaturisation of devices today

as technology advances is overwhelming. Devices such as microprocessors,

microelectromechanical systems (MEMS), nanoelectromechanical systems (NEMS),

microchannels and lab-on-chips come with high-density heat flux that needs quick heat removal for

2
efficient performance, stability and durability, which, from all indications, could be provided by

nanofluids for now [9-12]. The following are also emerging areas of applications of nanoparticles

and nanofluids: (i) they could be useful in medicine in the targeted treatment of malignant cells

without damaging healthy tissues, (ii) they could also be used in the biomedical field for drug

delivery for some special cases and (iii) they could also be used in surgery in order to increase the

chances of survival of patients [6, 13-16].

The two most important parameters in bringing about the efficiencies highlighted above are the

thermal conductivity and the viscosity of the nanofluids [17-20]. The thermal conductivity and heat

capacity of the fluid determine to a great extent the heat removal capacity of the fluid. The higher

the thermal conductivity, the more heat the fluid can remove from thermal systems. On the other

hand, viscosity is mostly important in systems that require flow because flow properties such as the

Reynolds number, heat transfer coefficients and pressure drop are very much dependent on the

viscosity [21]. Therefore, if the viscosity is very high, there will be a penalty on the pumping

power requirement to achieve the system’s target. In the recent past, much research progress has

been made on the thermal conductivity of nanofluids [22-32], which are a few of the copious works

that can be found on theoretical and experimental reviews of the thermal conductivity of

nanofluids. However, concerning the viscosity of nanofluids, very few theoretical models have

been developed based on the unique properties of nanoparticles [2, 33]. Most theoretical models

available were developed for the suspension of microparticles. Notably, there are many different

types of empirical models of nanofluids, but their usage is often limited to specific types of

nanofluids, nanoparticle sizes and volume fractions [34, 35]. The available review articles on

nanofluid viscosity is very scarce [36, 37], and they lack the appraisal into the existing numerical

works on nanofluid viscosity, which is part of the focus of this paper. Some other researchers have

made efforts in partly reviewing nanofluid viscosity, however, their efforts were not exhaustive as

their focus was mainly on the review of the thermal conductivity of nanofluids [38–41].

3
The mismatch between model and experimental results obtained from studies [42–44] on the

rheological behaviour of nanofluids is of great concern. Besides, [2, 45] have shown that despite

good agreement between experimental and theoretical results in certain cases, a wide range of

constitutive factors need to be incorporated into the models in order to account for the rheological

behaviour of the nanofluids in widely varying conditions [39]. Therefore, this work critically looks

into the viscosity of nanofluids from an analytical, empirical and numerical view, points out loose

ends in existing research and its methodologies and proposes an algorithm-based approach for the

selection of appropriate nanofluid viscosity model(s). This review is divided into five sections, viz.

1. Introduction, 2. Theoretical background of suspension rheology, 3. Experimental studies, 4.

Numerical studies and 5. Conclusion. Subsections under Section 1 are classical theoretical models,

which give a detailed account of colloidal suspension viscosity models that predate the invention of

nanofluids; new theoretical models, which are the models derived based on the factors that have

been identified as key players in nanofluid viscosity enhancement such as particle size, zeta

potential, pH and electrical double layer and empirical correlations, which are products of various

experimental studies and observations that have been carried out on nanofluid viscosity. Section 3

accounts for the methods that have been used in establishing nanofluid viscosity experimentally,

starting from the preparation of nanoparticles and nanofluids, characterisation and testing,

nanofluid stability and its markers, factors affecting nanofluid viscosity and a cursory look into

magnetic nanofluids for high-temperature lubrication application. The numerical aspect of

nanofluid viscosity enhancement under Section 4 brings the following to the fore: the application

of artificial neural network systems (ANNs), genetic algorithm and fuzzy logics for the prediction

of the viscosity of nanofluid, with a concluding part identifying areas for further development and

future research on nanofluids.

THEORETICAL BACKGROUND OF SUSPENSION RHEOLOGY

4
The rheology of colloidal suspension encompasses the study of the behaviours of suspension in

relation to whether it is Newtonian or non-Newtonian and thus a measure of its viscosity with

respect to shearing stress and shearing rate. There are a number of factors that affect colloids

suspensions (micro/nanoparticles in suspensions), which are: temperature, particle volume

concentration, particle size, particle size distribution (PSD), packing fraction, electrical double

layer (EDL), aspect ratio of particles, particle interaction, particle agglomeration, zeta potentials,

pH, nanolayer and magnetic properties of some particles [46]. As mentioned previously, the earlier

classical theoretical analyses were on microparticles in suspensions and a great deal of

simplification on the mechanisms of rheological behaviours was applied.

Classical Theoretical Models

Hypothetical analyses of the possible phenomena affecting the viscosity of nanofluids can be found

in the literature, though they are very limited when compared with the depth of the theoretical

models that can be found on thermal conductivity of nanofluids. However, several theoretical

studies have been conducted into the rheology of suspensions. The fundamental work by Einstein

[47] on infinite dilute suspensions of uncharged hard spheres based on the vorticity of the particle

shear field was the first available theoretical work on the viscosity of suspension that gave the

model in Eq. (1).

eff  o 1    
(1)

where eff is the effective viscosity of the suspension, o is the dynamic viscosity of the base fluid

and   is the intrinsic viscosity of the suspension. This linear equation is based on the assumed

absence of interaction between the particles, and the coefficient   is a function of shape of the

particle, which for hard spheres was given as 2.5. This model was stated to be valid for solid

volume concentration,   2% .

5
Numerous models were developed in efforts to extend the Einstein model to concentrated

suspensions [48–51] a few years after Einstein’s work. Contained in Table 1 is the comprehensive

list of other available classical models developed by various investigators [52–62] and are

applicable to the determination of the viscosity of solid-liquid suspension. These models differ

from one another, and no single model can predict data in the entire concentration range up to the

maximum concentration possible.

Contrary to the uncharged particles of Einstein [47], Smoluchowski [48] presented an effective

viscosity model for charged particles in electrolyte suspension given in Eq. (2). There was no

explanation of how this equation was achieved:

 
 1   DE  
2

eff  o 1  2.5 1  2    (2)
  k o a  2  
 

where  k is the specific conductivity of the electrolyte, a the radius of the solid particles, DE the

dielectric constants of the water and ζ the zeta potential of the particle with respect to the

electrolytic medium. There was no explanation of how this equation was achieved and this was

buttressed by researchers in the past [63, 64]. Based on experimental data, Bull in 1940 [64]

suggested that the effective viscosity of suspension of egg albumen varies with the square of the

electrophoretic mobility (a measure of zeta potential). However, the analytical works of

Smoluchowski [48] gave effective viscosity of suspension with the square of zeta potential as well,

unfortunately when this equation was applied to Bull’s experimental data, it exaggerated the

predictions. Therefore, Bull [64] proposed a simple model for effective viscosity at isoelectric

point to be:

s 
 0.0112 e (3)
 k

where  s is the specific viscosity,  e the electrophoretic mobility and k the specific

conductivity. Booth [63] in 1950 studied the overprediction (of viscosity of suspension with respect

6
to the effect of electroviscous force between particles and the suspending medium) made by

Smoluchowski’s model. Therefore, Booth made a quantitative recalculation of the electroviscous

force effect on effective viscosity, which predicted the data of Bull [64] with a good degree of

accuracy and the model is given as:

 
 
 e  
I

eff  o 1  2.5 1   bI 
   (4)
 
  kbT  

 1

where bI is the characteristics of electrolyte, e is the electronic charge on particles, k is the

Boltzmann constant and T is the absolute temperature.

In 1922, Jefferey [49] furthered the work of Einstein for suspensions that contain ellipsoidal

particles. Based on the principle of dissipation of energy, the model presented was not different

from Einstein’s Eq. (1). However, the intrinsic viscosity was provided with two limits (minimum

and maximum) for both prolate and oblate spheroids of different ellipticity of the meridian section.

Therefore, as the particles approach the sphere in shape, the difference in the limits of the intrinsic

viscosity diminishes and hence reduces to Einstein’s model. Ward and Whitmore [65]

experimented on microsphere-aqueous suspension in a bid to verify the Einstein equation. They

concluded that the intrinsic viscosity given by Einstein is a function of PSD ratio, which is

approximately 4.0 for an infinitely diluted suspension with PSD ratio 1:1 and approximately 1.9 for

PSD ratio exceeding 3:1. At the ratio of 1.5:1, the Einstein intrinsic value of 2.5 was obtained. This

was corroborated by the work of Vand [52]. Williams [66] also concluded that PSD is a key

parameter affecting the viscosity of solid-liquid suspension after experimenting with different size

(4, 8 and 12 µm) categories of glass spheres in an ethylene glycol-water mixture. He tried to fit the

data obtained into the equations of Mooney [53] and Roscoe [54] with relatively good success.

Maron and Fok [67] held that the duo of Mooney and Roscoe models did not satisfactorily predict

the experimental data of Williams [66], hence tried to fit the data with models of Maron and Fok

[67] that had been successfully tested with lattices and latex mixtures. After treating the

7
experimental data with the least-squares method to obtain model constants, they clearly showed

that if the intrinsic viscosity of Einstein’s model were to be between 3.15-3.35, the equations of

Mooney [53] and Roscoe [54] would have given a perfect fit to Williams’s data.

Applying a different viewpoint, Batchelor [55] considered the influence of interparticle interactions

to obtain the model in Eq. (5), for the relative viscosity of solid-liquid suspension for a case of

volume fraction,   4% . Within the limits of a very low particle concentration, this model

approaches Eq. (1), which means Batchelor’s model does not differ from Einstein’s model, i.e. at

low volume concentration, the assumption of non-interaction of particles as assumed in [47] is also

inherently considered and this is an ideal situation.


eff  o 1     kH   
2
 (5)

where k H is the Huggin’s coefficient known also as the interaction parameter, this coefficient

accounts for interparticle interaction as opposed to hydrodynamic effects [68]. The semi-empirical

relationship proposed by Krieger and Dougherty [56] for shear viscosity covering the full spectrum

of particle volume concentrations is expressed as:

 m
  
eff  o 1  
 m  (6)

where m is the maximum particle volume fraction at which flow can still occur (i.e. the

concentration at which the relative viscosity approaches infinity asymptotically), the intrinsic

viscosity [  ] was given as 2.5 for monodispersed suspensions of hard spheres. However, in

practical situations, particles are polydisperse in nature. Hence the assumption made by Krieger

and Dougherty [56] is not valid for all particulate suspensions and this has been accentuated with

underpredictions of this model when applied to viscosity data, for example, when it was applied to

Al2O3-water nanofluid data [69].

8
Currently, there is a lack of unified models that can be used to predict the viscosity of colloids.

Most of the available classical models are built around particle volume concentration and when

tested, they all give different predictions as depicted in Fig. 1. The insets of Fig. 1 show the degree

of variations among the models. There is no clear-cut phenomenon to explain the erratic nature of

nanofluid viscosity data presented by different investigators. However, using these models to

predict the recent experimental data as depicted in Fig. 2 and Fig. 3 (for Al2O3-water and TiO2-

water respectively) shows their inability to accurately predict the suspension’s viscosity.

Generally, the discrepancies in reported viscosity data have mostly been ascribed to agglomeration

formation. Chen et al. [68], based on this widespread assertions, extended the theoretical work of

Krieger and Dougherty [56], which was based on packing fraction of monodispersed particles

without agglomeration. Chen et al. [68] assumed that if particle agglomerates were spherical, the

sphere would be of different sizes. Thus they derived a modified Krieger and Dougherty equation

as presented in Eq. (7), based on maximum packing fraction of agglomerates and the fractal index

of the agglomerates, which is an indication of degree of variation in the packing fraction from the

centre of the agglomerates to the outer edge.

 m
  
eff  o 1  a  (7)
 m 

a is given by a   ma , where ma is the packing fraction of the aggregates. The viscosity was

assumed to follow a power law with a fractal index, D. Consequently, a becomes a    aa a 3 D ,

where aa a is the ratio of effective radii of aggregates and primary nanoparticles.

In an attempt to bring together the two separate views, namely that (i) PSD affects the viscosity of

suspension and further that (ii) agglomeration, which is a function of interaction between particles,

affects the viscosity of suspension, Farris [70] suggested that agglomeration alone could not

describe the evolution of viscosity of nanofluids and that PSD was seen to play a role in the

viscosity trend. Therefore, if the PSD is discrete, the global relative viscosity of the non-interacting

9
monodispersed system of particles in suspensions may be calculated as the product of each

independent viscosity as shown in Eq. (8) [70].

 11   22    zz 

r        (8a)
 o   o   o 

z
In general form r  r i  (8b)
i 1

where i is the zth class corresponding particle fraction and z stands for the different average

particle sizes contained in the distributions. The viscosity of each monodispersed suspension can

also be related to the maximum particle volume fraction, m with the proposed model in [71] given

as:

2
   /  
r i   1  0.75  i m   (9)
  1  i m  

Chong et al. [71] determined experimentally the maximum particle volume fraction for a

monodispersed system of glass beads sized in microns using a specially developed orifice

viscometer. They reported m for their experiment as 0.605 and further proved that plotting the

r r  1 vs  and extrapolating up to the point where the two axes variables meet will give the

m for any suspension. Storms et al. [72] expanded the work of Chong et al. [71] to understand the

effect of plydisperity (size ratio and % volume fraction of smaller particles) on the viscosity of

suspension. It was found that the viscosity is dependent on the size ratio and % volume fraction of

the small particles present. The model in Eq. (10) predicted the experimental data of Chong et al.

[71] with good accuracy:

3.3m
 Ri 
r i   1   (10)
 1  i m 

10
where R is an adjustable parameter and varies from 0.7 to 1.25 depending on the size distribution.

If m is taken as 0.605 as Chong et al. [71] proposed, the exponent in Eq. (10) becomes 2. It should

be noted that Eqs. (9) and (10) are modified versions of the Eilers equation given below.

2
 2.5 
r  1   (11)
 2 1   m  

Determining m for monodispersed and bidispersed suspension, Dames et al. [73] applied an

empirical model in Eq. (12) to determine the maximum packing fraction for particle sizes Rlarge =

270 nm and Rsmall = 80 nm:

  D5  
m  z  z   m 
mono  0.271 D  
  1  (12)

N d i i
x

where z  1  1   m 
mono z
and Dx  i 1
z

N d
i 1
i i
x 1

where mmono is the maximum packing fraction of a monodispersed system and it is taken as 0.63, z

is the number of modal suspensions (mono, bi or multi), Dx is the x-th moment of particle size

distribution, i.e. D1 is the number average of the particle diameter. For a multidispersed mixture of

spherical particles, the technique given in Muralidaharan and Runkana [74] and Servais et al. [75]

can be used to determine m from the minimum value of Pi :

m  min  Pi  (13)

where Pi is the packing fraction of each class size i , calculated based on the following:

n
Pi    ij j (14)
j 1

 ij is the binary packing coefficient of classes i and j, and  j is the volume fraction of the class j.

The procedures for calculating these two quantities are detailed in [74, 75].

11
It should be noted that all of these earlier works were done before the invention of nanofluids and

as such were basically on the suspension of microsolids and rigid spheres in fluid medium.

However, knowledge and ideas have been borrowed from these past works for the recent analyses

of the thermophysical properties of nanofluids. Again, some of these models were first published

more than 80 years ago. Therefore, the useability of these classical models on nanofluids is a

subject that should be critically analysed [76] because many factors that affect colloids in

suspensions have been greatly oversimplified in order to achieve a presentable solution.

New Theoretical Models

Researchers have tried to predict the viscosity of nanofluids using the classical models on viscosity

of suspension without success. These models all predate the invention of nanofluids. Therefore,

some very salient characteristics such as nanolayer, pH, electrical double layer (EDL), zeta

potentials, temperature, capping layer, interparticle spacing and particle magnetic properties that

influence the thermal properties of suspensions were not considered.

Chen et al. [68], based on an earlier work, considered an important factor (agglomeration) because

it affected the viscosity of nanofluids. After substituting some empirical data that described the

extremes of nanoparticle agglomeration, the viscosity of nanofluids given in Eq. (15) was derived.

1.5125
   aa  
1.2

r  1 
 0.605  a  
(15)
 

Hosseini et al. [33] obtained a new dimensionless model for predicting the viscosity of nanofluids.

The relative viscosity was a function of formulated dimensionless groups, which contains the

following parameters: (i) viscosity of the base fluid, (ii) the hydrodynamic volume fraction of

nanoparticles, (iii) diameter of a nanoparticle, (iv) thickness of the capping layer, and (v)

temperature as dimensionless groups  1 ,  2 ,  3 and  4 respectively as presented in Table 2. The

model (Eq. 16) is the result of the combination of the dimensionless groups:

12
  T   d p 
nf  o .exp  m       h      (16)
  To   1  r  

where  nf is the viscosity of the nanofluids, o is the viscosity of the base fluid, h is the

hydrodynamic volume fraction of nanoparticles, d p is nanoparticles diameter, r is the thickness of

the capping layer, To is a reference temperature taken to be 20 oC, and T is the measured

temperature. In this model, m is referred to as system property constant, which is a function of

types of nanoparticles, types of base fluids and the interactions between them.  , ,  are

empirical constants obtainable from the experimental data. However, there was no indication of

how these constants were derived and this makes the testing of these models against similar or

other nanofluids a problem. The model was also tested with limited samples of Al2O3-water-based

nanofluids, although it claimed good agreement with the experimental data tested.

Recent literature revealed that the theoretical analysis of the effective viscosity of nanofluids can

be approached as either a single-phase problem or as a two-phase problem. Masoumi et al. [2]

analysed the dispersion of nanoparticles in a fluid medium as a two-phase problem and considered

five parameters as affecting the effective nanofluid viscosity. The following parameters:

nanoparticle size, temperature, nanoparticle density, nanoparticle volume fraction in the fluid

medium and fluid physical properties were considered. Using the Brownian velocity relation

presented by Prasher et al. [77] in Eq. (17), the effective viscosity was derived as presented in Eq.

(18). A creeping flow assumption around the spherical nanoparticle was used in their derivation

and they introduced a correction factor, also based on very limited data to take care of the

simplification of their assumption. The model was tested in predicting effective viscosity of CuO-

H2O, CuO-EG, TiO2-EG, CuO-EG/H2O and Al2O3-H2O nanofluids and according to the results

presented, there was an acceptable level of agreement between the model and the experimental data

used.

13
 1 18kbT
VB  (17)
d p  p d p

 pVB d p2
nf  o  (18)
72C1

where kb is the Boltzmann constant, T is the temperature,  p is the particle density,  is the

distance between the centres of particles, VB is the Brownian velocity, C1 is the correction factor,

and d p is the particle diameter. VB ,  and the constant were C1 defined as:


3 dp (19)
6

C1  o1  c1d p  c2     c3d p  c4  (20)

constants c1  c4 are obtainable from experimental data. It should be stated here that this is one of

the very few theoretical analyses of nanofluid viscosity existing in the literature. However, the

required procedures as described by the authors to obtain this set of constants will not allow for

reproducibility of these constants and hence C1 may be difficult to calculate for other nanofluids

different from those in their work. We believe a better presentation of these constants can be made,

in fact, to our knowledge, the model has not been cited much in comparison with experimental data

by other investigators since its publication.

Empirical Models

Ward in [78] recommended that experimental results should be expressed in the form of Eq. (21) to

allow easy comparison with theoretical models. He further noted that the intrinsic viscosity should

be determined experimentally, because it is difficult to evaluate the intrinsic viscosity from the

power of three and above theoretically [79].


eff  o 1        2     3     4  ..
2 3 4
 (21)

14
Cheng and Law (CL) [80] reanalysed effective viscosity of suspensions based on Einstein’s model

to provide an exponential formula for the effective viscosity of nanofluids for volume fractions

higher than Einstein’s concentration regime. The CL model (Eq. 22), though similar to the general

model expression given by Batchelor [55] and Lundgren [57], however, provided the coefficient of

volume fraction up to the power of five. When compared with the experimental data reported by

Ward in [78], they are in close agreement.

 2
 35 5  2  105 35 5 2 3 
3

1  2.5              
  8 4   16 8 12  
 4 
  1155 935 235 2 5 3  4 
eff  o           (22)
  128 96 96 48  
 
  3003  1125   1465  2  95  3  1  4   5  ...... 
  256 64 192 96 48 
 
 

where  is the diffusion exponent. Avsec and Oblak [45] emphasise that Ward’s model as

presented in Eq. (23) is of little importance to nanoscale viscosity (nanoviscosity) and presented a

new model (Eq. 24) with a simple twist to Ward’s expression. This was derived using statistical

mechanics owing to the possibility of modelling particulate interaction (nanolayer interaction

effect) with statistical mechanics.


eff  o 1  2.5   2.5  2   2.5  3   2.5  4  ...
2 3 4
 (23)


eff  o 1  2.5eff   2.5 eff 2   2.5 eff 3   2.5 eff 4  ...
2 3 4
 (24)

3
 h
where eff   1   (25)
 a

eff is the effective volume fraction, h is the thickness of the nanolayer and a is the particle radius.

Apart from the theoretical models presented above, most of the available models for the

determination of nanofluid viscosity are correlations from very limited experimental data. These

models are not hybrid because they are in most cases developed from experimental data with a

15
confined volume fraction of nanoparticles, a few nanoparticle types (at most three for a model), a

small spectrum of the nanoparticle size and mostly spherical or assumed spherical in shape [81–

86].

It has been observed by investigators through experimental studies that the temperature of medium

of study, volume fraction, shear rate and size of nanoparticles affect the effective viscosity of

nanofluids [36, 87, 88]. However, an exhaustive examination of the existing empirical models

shows that in the majority of the correlations, researchers refrained from providing the effect of

temperature, shear rate and size of nanoparticles as they affect nanofluid viscosity in their

correlations.

Nguyen et al. [87], after a comprehensive investigation of the dynamic viscosity of alumina-water

nanofluids considering different nanoparticle volume fractions, sizes and temperatures, only

provide individual correlation equations for nanofluid viscosity based on the volume fraction and

temperature respectively. In fact, the models with temperature can only predict for 1%-4% volume

fraction, which was even inadequate for their experimental data. Similarly, Vakili-Nezhaad and

Dorany [89] provided two empirical models for the same nanofluids (SWCNTs/lube oil) based on

volume fraction and temperature. The correlations are polynomial function of volume fraction and

temperature as given below:

nf  o 1  1.59  16.36 2  50.4 3  (26)

nf  o 1048  30.3T  0.2T 2  (27)

Eqs. (26) and (27) are valid for 0.01-0.2% volume fraction and 25-100 oC temperature respectively.

Rashin and Hemalatha [35] on the viscosity of CuO-coconut oil also proposed two separate models

to predict the mass fraction and temperature dependence of their experimental data. The mass

fraction model, similar to Batchelor’s equation [55], is presented alongside the temperature model

in Eqs. (28) and (29):

nf  o 1  A  B 2  (28)

16
 nf  Ce0.03T (29)

where  is the mass fraction of nanoparticle to base fluid, T is the temperature in Kelvin, A, B and

C are parameters from regression analysis and unique to 20 nm CuO-coconut oil nanofluids at a

temperature range of 35-55 oC. Many other researchers just present their model as a linear or

polynomial function of the volume fraction without considering the effect of at least temperature

[90–94].

Heyhat et al. [21] presented a volume fraction exponential correlation relating the viscosity of

alumina-water nanofluid at volume fraction of 0.1-2%. The correlation (Eq. 30) averaged the effect

of volume fraction over the temperature range experimented (20-60 oC):

nf  5.989 
 Exp   (30)
o  0.278   

In 2012, Suganthi and Rajan [95] proposed a modified form of Einstein’s equation (Eq. 31a) by

replacing the volume fraction with agglomerate volume fraction similar to Chen et al. [68]. This

was done in order to account for the effect of agglomeration on the viscosity of nanofluid.

nf  o 1  2.5a  (31a)

where a is related to  using:

3 D
a  Da 
  (31b)
  d p 

Generally, in nanofluids, agglomerates are of different sizes. Therefore, Eq. (31b) can be rewritten

in a more broad form:

 Da 
3 D
 D
N 
3 D

     i  ai   (31c)
i 
d 
 dp    p  

where i is the mass fraction of the aggregate i, Dai is the diameter of the aggregate i, N is the

number of aggregates present and D is the fractal dimension. The diameter of the nanoparticle

17
(ZnO) used is 35-40 nm. As the volume fraction increases up to 2%, the authors proposed a

modified Batchelor’s equation [55] as shown in Eq. (32) to take care of the effect of particle-

particle interactions:

nf  o 1  2.5a  6.1a2  (32)

Recently, Suganthi et al. [96] proposed a temperature-based power law correlation (Eq. 33) to

relate the effect of temperature on the viscosity of ZnO-propylene glycol (PG) nanofluid. Their

previous models (Eqs. 31 and 32) could not be applied here, probably because the trend of ZnO-PG

with increase in volume fraction is directly opposite to the trend in their experiment on ZnO-water

nanofluid [95]:

nf  A  B (33)

where  is in degree Celsius, A and B are empirical constants, and are different for different

volume fractions. Singh et al. [97] offered a correction based on the Arrhenius functional form (Eq.

34) to predict the temperature dependence of 170 nm SiC-water-based nanofluid at volume fraction

of 1.8, 3.7 and 7.4%:

 Ea 
nf   ,T exp   (34)
 Rg T
 

where ,T is the viscosity at “infinite temperature”, Ea is the activation energy to viscous flow, Rg

is the universal gas constant and T is the temperature in Kelvin. The viscosity at infinite

temperature and the activation energy can be obtained from experimental data, using the

logarithmic form of the Arrhenius equation [98]. Also, Abareshi et al. [43] were able to fit their

experimental data at different volume fractions into the Vogel-Fulcher-Tammann (VFT) equation

[99], which took care of the temperature effect on nanofluid viscosity alone as shown below.

 B 
nf  
T   Ae T T o 

o (35)

18
Eq. (35) was fitted for the two limiting volume fractions to produce two sets of the empirical

constants A, B and To. The problem with generating correlation through regression analysis of a

single parameter such as temperature is that when large numbers of volume fractions are involved,

the empirical constants become huge and untidy. This is also dependent on the number of constants

in the proposed model. For instance, Zyla et al. [100] proposed a nine-order polynomial correlation

to fit the viscosity of Y2O3-diethylene glycol (DIEG). The variable in their correlation is the

temperature of the nanofluid. The concentration studied was in five levels (5:5:25%). Therefore,

according to their model Eq. (36), the empirical constants totalled 50.

r  a0  a1T  a2T 2  a3T 3  a4T 4  a5T 5  a6T 6  a7T 7  a8T 8  a9T 9 (36a)

In a more general form:

9
r   aiT i (36b)
i 0

where ai ' s are the experimental constants and T is the temperature in degree Celsius.

Kulkarni et al. [101] gave a correlation, which is one of the few empirical works that take care of

both the temperature and volume fraction effect on the nanofluid viscosity. Another of such is the

work of Namburu et al. [34], in which they tried to fit their experimental data to existing equations

in the literature. However, the failure of the exercise spurred a new correlation given in Eq. (37),

which considered both temperature and volume fraction effects.

Log  nf   Ae  BT (37a)

Eq. (37a) is an empirical model with a correlation coefficient R2 = 0.99, developed for CuO-EG

nanofluids with volume concentration of CuO ≤ 6.12%. Constants A and B were calculated with

the correlations below:

A  1.8375 2  29.643  165.56 


2 
(37b)
B  4 10   110   1.86 10 
6 2 3


19
where RA  0.987 and RB  0.988 . Nguyen’s correlation [87] performed better than most of the
2 2

classical models. However, Yiamsawas et al. [102] showed that Nguyen’s correlations under

predicted their alumina-EG/water experimental data, because Nguyen’s correlation considers

separately the volume fraction and temperature effects on the nanofluid viscosity. Consequently,

the correlation below was developed to predict the viscosity of titania-EG/water (20:80) and

alumina-EG/water (20:80) with volume fraction range of 1-4%. The experiments were conducted

between 15-60 oC, and the diameters of the titanium oxide and aluminium oxide used were 21 and

120 nm respectively.

nf  A BT C oE (38)

where o is calculated based on this expression; o  2.3775  0.0461T  0.0003T 2 , A, B, C and E

are empirical constants obtained from regression analysis. Lately, Hemmat Esfe and Saedodin

[103] offered a two-variable correlation (namely temperature and volume fraction) for ZnO-EG

nanofluids. The correlation (Eq. 39) has an average deviation of 2% from the experimental data,

and the stability region over which it was tested is 25 ≤ T ≤50oC and 0.25 ≤ ≤ 5%:

nf 5.49 0.00001359T 

 0.0303ln T 
2

 0.9118e (39)
o

Azmi et al. [104] offered a water-based correlation based on the combined effects of volume

fraction, temperature and nanoparticle size on the effective viscosity of nanofluids. Using data

available in the literature on Al2O3, CuO and SiC with particle size ranging from 20-170 nm and

volume fraction ≤ 4%, they proposed the following correlation:

  
nf     Tnf   dp 
 C1 1   1   1   (40)
o  100   70   170 

20
where C1 is empirical constant and the exponents  ,  and  are 11.3, 0.038 and 0.061

respectively. To test the performance of this model, they carried out new experiments on water-

based nanofluids of Al2O3, ZnO and TiO2. Generally, the model is valid for water-based nanofluids

of Al2O3, CuO, Sic, ZnO and TiO2 with particle diameter between 20-170 nm and volume fraction

 ≤ 4%. Khanafer and Vafai [105] also developed a three-parameter correlation to predict the

viscosity of Al2O3-water nanofluid. Their correlation shown below is valid for volume fraction

between 1-9%, temperature between 20-70 oC and nanoparticle diameter between 13-131 nm.

28.837 2  2 3 (41)
nf  0.4491   0.574  0.1634 2  23.053 2  0.0132 3  2354.735 3  23.498 2  3.0185 2
T T T dp dp

The model above predicted the viscosity changes with temperature of some data in the literature

with good accuracy. However, the performance of the model was not tested against volume

fraction increase.

Numerous studies have shown that the addition of nanoparticles to a Newtonian base fluid

sometimes turns the fluid to non-Newtonian. An example of this is the recent works of Yu et al.

[106] on aluminium nitride nanofluid and Halelfadl et al. [107], and Aladag et al. [108] on alumina

and carbon nanotube water-based nanofluids. In most situations, despite the non-Newtonian nature

of nanofluids, researchers will only provide correlations considering temperature and volume

fraction or both. According to Syam Sundar et al. [109], it is important to study the behaviour of

nanofluids with respect to change in shear rate, therefore empirical correlations should be designed

taking cognisance of this parameter. Hernández Battez et al. [88] investigated the rheology of ZnO

and ZrO2 suspended in Newtonian polyalphaolefin (PAO 6) base fluid. At low shear rates (0-700 s-
1
), all nanofluid samples behaved as Newtonian fluid; however, at higher shear rates (106 -107 s-1),

non-Newtonian shear thinning characteristics with varying trends were observed. This suggests that

the effect of shear rate is significant in the characterisation of nanofluid viscosity. Therefore, the

following correlations were proposed for the two types of nanofluids (ZnO and ZrO2) respectively:

21
 r (cp)  52.80  9.76 107   0.172  0.912T  1.02 108  T  4.24 103T 2 (42)

r (cp)  53.78  9.25 107   0.202  0.937T  9.65 109  T  4.39 103T 2 (43)

In the equations above,  is the shear rate,  is the mass fraction and T is the temperature. The

correlations factored in the effect of shear rate using the least-squares approach with correlation

coefficients R2 > 0.995 in both cases. Earlier, Phuoc and Massoudi [110] noted that the addition of

nanoparticles created yield stress within the Fe2O3-water/dispersants (polyvinylpyrrolidone, PVP

and polyethylene oxide, PEO) nanofluids. They described the dependence of the yield stress on

volume concentrations as  o   en . Using the Casson equation (for flocculated suspended

particles) described by  1 2   o1 2  1 2 1 2 to determine the suspension viscosity at an infinite shear

rate for their base fluid plus dispersants, they found that the viscosity values predicted with

Casson’s equation were two orders of magnitude lower than the viscosity of the base fluid.

Moreover, the model does not consider the effect of particle concentration. Therefore, the

following correlation was proposed to characterise the combined effects of shear rate and volume

fraction on the viscosity of Fe2O3-water/(PVP and PEO) nanofluids:

  en  2   en  2 
1 1
1
nf          2  
2
(44)
      


The data used for the correlation in Eq. (44) were taken at 25 oC, where  is the intrinsic viscosity

at infinite shear rate,  and n are empirical constants dependent on the intercept of the plot of

   against    for all the volume fractions of nanofluids. A summarised description of the
12 12
nf

above empirical models and other empirical models [32, 44, 81–83, 85–87, 90–92, 94, 101, 109,

111–114] with regards to the concentration regime, particle size, temperature and remarks is

presented in Table 3.

22
EXPERIMENTAL STUDIES

There are copious experimental studies on nanofluid thermal properties and their behaviours. These

studies mainly cut across the determination of nanofluid thermal conductivity [115–117],

convective heat transfer [118, 119], performance in heat pipes and microchannels [120–122], its

effectiveness in solar heaters and solar-related devices [123–126] and its behaviour in car engine

radiators [76, 127]. Few of these works focused on the viscosity of nanofluids and recently there

has been some new development in the experimental field, as researchers are now investigating the

thermal diffusivity and electrical conductivity of nanofluids and how these properties affect

thermal properties [128, 129]. For consistency in this review, the following are the experimental

works done on the viscosity of nanofluids:

Methods of Preparation of Nanoparticles and Nanofluids

The preparation of nanofluids can be classified into two groups: (i) preparation of dry

nanostructures and subsequent dispersion of the nanostructures in the base fluids, and (ii) an

infused method of preparing nanostructures in its intended base fluid medium. The first approach is

called the two-step method, in which case, nanostructure preparation is the first step to the

development of nanofluids. Nanostructures of sizes ranging from 1nm-100 nm are desired for

nanofluids [17, 130] and a wide range of nanostructures exist today, ranging from nanowires,

nanorods, nanofibres, nanocylinders, nanograins to nanoparticles [98, 131, 132], which can either

be developed from the smallest unit of matter (known as the bottom-up method) or fractured down

from bigger lumps to nano-sized particles (known as the top-down method) [133–135]. The most

common methods of nanostructures synthesis are broadly classified into physical and chemical

methods. Some of the physical methods are pulsed laser ablation [136–138], laser deposition and

matrix assisted pulsed laser evaporation (MAPPLE) [130, 139], and ball milling [133–135], while

the chemical methods are chemical precipitation [140–142], sonoelectrochemical synthesis [143–

148], spray pyrolysis [149, 150], chemical vapour deposition [151–153] and thermal

decomposition [154]. Details on other methods under these categories can be found in the past

23
publications [38, 155, 156]. After the synthesis of the desired nanostructure, the nanostructures are

dispersed in any intended base fluid medium using assisted dispersion with magnetic stirrer, high

shear homogeniser, high-pressure homogeniser and/or ultrasonication (probes and baths) [157,

158]. The production of nanofluids by this first method is very popular and by far the most

economical method which has been transformed to industrial scale of production. However, there

is a problem of stability with nanofluid prepared using this method. The second method is called

the single-step method because the process of preparing nanostructures (nanoparticles in most

cases) is infused into preparation of the nanofluids without any intermediate product, i.e. no dry

nanoparticles are produced, rather the nanofluids are produced in the form of a continuous process

[159, 160]. Nanofluids from this method are by far more stable, but it is very costly, which is one

of the hindrances why it is yet to reach commercialisation.

Worthy of mention is the biological method of synthesis of metallic nanoparticles. Using different

biological substances such as algae [161], fungi [162, 163], plant extracts [164–168], bacteria

[169–171], yeast [172–175] and marine sponge [176] have been used to synthesise two metallic

nanoparticles, which are silver (Ag) and/or gold (Au). Thakkar et al. [177] established in their work

the need for eco-friendly metallic nanoparticle synthesis, vis-à-vis energy saving, poisonous gas

reduction and decline in the usage of toxic compounds. Therefore, a comprehensive usage of

different micro-organisms for the production of non-toxic metallic nanoparticles was highlighted.

Similarly, Elumalai et al. [164] and Krishnaraj et al. [165] synthesised silver nanoparticles from

different leave extracts and both studied their antibacterial activities. Table 4 gives an overview of

the classifications and different methods that have been used in the synthesis of nanostructures.

Nanofluid Stability

The versatility of nanofluids cuts across all phases of science and technology in human

developments and this is what makes it one of the most popular and fastest-growing research

entities in the world today. This versatility has also faced many challenges, most of which have

24
been met to a reasonable degree. However, the problem of the stability of nanofluids still persists.

Like stability in many other engineering systems, a stable nanofluid is when nanoparticles are

continually in their Brownian motions without cohesion and devoid of flocculation, agglomeration

and ultimately sedimentation. It may be said that at very low nanoparticle volume concentration,

stability does not pose a big threat in the face of the present methods of preparation of nanofluids.

However, in the case of increasing volume concentration as desired by many applications,

instability becomes very pronounced vis-à-vis flocculation, agglomeration, settling down and

ultimately phase separation. Though, when first prepared with any combination of the different

dispersion methods available [157] (especially the two-step method), it displays good dispersion,

mostly because of the high shear, pressure or energy impacted into the homogenisation process.

However, with time, agglomeration formation occurs [178] as shown in Fig. 4. The time difference

between to when a homogeneous suspension is formed and t3 when full-phase separation occurs is a

function of compositions, i.e. particle volume concentration, nanoparticle type, type and

concentration of surfactant, shape of particle, type of system (stationary or dynamic), method of

preparation, temperature of the suspension and density difference between the nanoparticles and

base fluid. For example, exploring three different dispersion methods, namely ultrasonication, ball

milling and high-pressure homogenisation (HPH), Fedele et al. [157] prepared deionized water (DI-

water) based nanofluids of CuO, TiO2 and single-wall carbon nanohorns (SWCNHs) to study their

stability. With sedimentation rate as stability marker, and applying static and shake dynamics (i.e.

samples were prepared and divided into two, first sets were kept stationary while the second sets

were shaken before measurement), they measured using dynamic light scattering to monitor the

variation in size of dispersed nanoparticles and SWCHNs over a period of 30 days. From their

findings, CuO and TiO2 dispersed using ball milling sedimented four days after preparation

whereas the same nanoparticles dispersed through ultrasonication lasted more than 15 days in

suspension. On addition of surfactant to the unstable TiO2-water nanofluids at varying %weight,

25
the nanofluid was made stable for more than 30 days without any sign of agglomeration or

precipitation.

Zamzamia et al. [158] formulated 90 days stable nanofluids of Al2O3 and CuO in ethylene glycol

and tested both in double-pipe and plate heat exchangers. Also, employing wet ball milling, Samal

et al. [179] produced and dispersed nanoparticles of Al-Cu alloy in DI-water to study its stability at

different pH. By measuring the zeta potential of the nanofluid as the pH is varied, their results

showed that pH is a major driver of stability of the nanofluid, with or without the addition of

surface active agents (surfactant). This is because pH modification has a direct link with the

electrostatic condition of the interface between the suspended particles and the fluid medium. In

another pH influence investigation, Wang et al. [180] systematically prepared optimised Al2O3-

water and Cu-water nanofluids by the two-step method using the zeta potential, nanoparticle size

and absorbency of the nanofluids as stability indicators. They varied the pH and surfactant

concentration (in this case, sodium dodecylbenzene sulfonate (SDS)) to synthesise a stable

nanofluid as indicated by the measured values of zeta potential and absorbency. However, at the

optimised pH and surfactant concentration that depicted stable nanofluids from the above-measured

parameters, the measured size of nanoparticles in suspension was more than the starting materials

by a factor of 7.4 for Al2O3 and 7.5 for Cu. Although, the sizes were the minimum considering the

sizes reported with pH and surfactant variations, the sizes also coincided with the optimised pH and

surfactant concentration. Because the measured size was bigger than the size of the starting

materials, one of two things could be inferred here, namely that the reported starting materials were

not actually the size reported or there was agglomeration within the nanofluids prepared as the time

given for sonication might not have been enough (in this case one hour).

In stability characterisation, the key marker of stability from the available systematic experimental

investigations [157–159, 179, 180] can be narrowed down to the following: visual inspection for

sedimentation, sedimentation rate measurement, turbidity, zeta potential, absorbency of nanofluids,

26
transmittance and size measurement against time after preparation (for detecting agglomeration

and/or reduction in nanoparticle population). Some researchers have also deduced that rheological

characteristics, i.e. Newtonian, signify stable nanofluids and non-Newtonian characteristics signify

otherwise [181]. All these characterisation procedures have their deficiencies, which are still

preventing the report of either qualitative results or a report that represents the actual situation that

is encountered in real-life application. For instance, visual observation of sedimentation is relative

to the eyes and as such lacks substance in reporting the stability of nanofluids. However, it can be

used as a secondary means of characterisation (i.e. to back up a quantitative report on stability).

Also, the quantitative approaches available such as turbidity measurement, zeta potential,

absorbency and nanoparticle size measurement all have restrictions. Either the volume

concentration must be in dilute regime (usually ≤ 1%), which is not always the case in real-life

applications or the opacity of the nanofluids is disrupting the proper measurement, in which case it

must be diluted as well. The idea of dilute concentration of samples in order to measure any of

these stability markers may not represent what is obtainable in the industry where a concentration

of up to 10% might be required for application in heat exchangers and other engineering problems.

Moreover, some nanoparticles are opaque in nature, e.g. Fe2O3, Fe3O4 Cu and CuO.

Experimental Set-ups

Measuring nanofluid viscosity seemingly looks very simple when the scale and sensitivity of

equipment requirement are compared with other aspects of the thermal properties of nanofluids.

Nevertheless, it is important to state that a good knowledge of the design of experiment and control

of variables to stability is required to be able to measure and report accurate data. At least five

types of viscometers with different principles of operations have been used in viscosity

measurements, viz. capillary tube viscometer [42, 94], vibro-viscometer [17, 182], rotating

viscometer which includes cone-plate, flat plate and concentric geometries [18, 32, 43, 44, 68, 76,

81, 82, 85, 86, 183–187], falling ball/falling piston viscometer [87] and cup-type viscometer.

27
Chen et al. [68] formulated an ethylene-glycol-based TiO2 nanofluid from dry nanoparticles using

the two-step method and in the absence of dispersant. After sonication of the nanofluid samples for

20 h each, samples were subjected to a rheological test using Bolin CVO rheometer that works

based on the principle of controlled shear stress. Al2O3-EG/water nanofluids were prepared and

sonicated in the presence of oleic acid as the surfactant, for 3 h. The suspension was further

agitated magnetically for 1 h to obtain a uniform homogenisation after which a rheological test was

performed between the temperatures of 20 oC-100 oC using the Brookfield programmable

viscometer (model: LVDV-II-Pro), which works by controlling the spindle shearing rate [76].

Chandrasekar et al. [32], with a similar set-up, measured the viscosity of Al2O3-water without

dispersant and sonicated for a period of 6 h. Table 5 shows various pieces of equipment that have

been used in the measurement of viscosity of different nanofluids with their measurement

principle. In a rather different set-up, Lee et al. [182] used a vibro-viscometer operating on a

constant resonance frequency to determine the viscosity of aqueous SiC nanofluids.

Parameters Involved in the Effective Viscosity of Nanofluids

Temperature

The reduction in the viscosity of conventional heat transfer fluids such as water, EG, PG, PAO,

engine oil and grease is an established phenomenon. The rate at which this occurs is commensurate

with the intrinsic properties of the fluid (intermolecular bond strength) [182]. Nanofluids, which

are a better replacement for these conventional fluids, have hardly been in use for two decades and

have drawn the attention of many in the industry, research centres and academia. In order to

understand and maximise the potential of these superfluids, a number of investigations into their

behaviours at different elevated temperatures have been carried out. Heating of fluids generally

supplies the molecules of the fluid with higher energy. This increase in energy contributes to

increased random motion and weakening of intermolecular forces holding the fluid molecules. This

phenomena result in a reduced resistance of the fluid to shearing flow, and by implication, a

reduction in viscosity is experienced.

28
It is important to note that the behaviour of the viscosity of nanofluids to change in temperature

does not in any way differ from the behaviour of the conventional fluid as stated in the paragraph

above. Kumaresan and Velraj [185] presented the relationship between temperature and the

viscosity of MWCNT-water/EG-based nanofluid at 0.15%, 0.30% and 0.45% particle volume

fractions. An increase in viscosity of the nanofluid at temperature below 25 oC was initially

observed, however, further increase in the temperature witnessed corresponding reduction in

viscosity of their nanofluid samples. In another recent report by Aladag et al. [108], CNT-water-

based nanofluid was investigated for its rheological properties at low temperature (2-10 oC). The

characteristic of the measured viscosity with respect to temperature increase was not different from

the behaviour of conventional heat transfer fluid as widely reported in literature. However, one

may conclude that the addition of surfactant in the experiment carried out by Kumaresan and

Velraj [185] might probably be responsible for the initial behaviour of the nanofluid.

With a much different type of nanofluid, Sahoo et al. [86] considered Al2O3-EG/water nanofluid

for its rheological characteristics. They found that it exhibited non-Newtonian behaviour at a very

low temperature. This behaviour specifically fitted into the characteristics of a Bingham plastic and

was more pronounced as the concentration of the nanoparticles increased. Nevertheless, as

temperature increased across the volume fraction investigated, the viscosity of the nanofluid

decreased exponentially. Similar results were obtained by Syam Sundar et al. [188] when they

investigated the viscosity of nanofluids synthesised from magnetic Fe3O2 and EG-water mixture.

The temperature of their investigation started from 0 to 50 oC at maximum volume fraction of

1.0%. Varying the % weight composition of the EG-water mixture (60:40, 40:60 and 20:80) did not

have any impact on the viscosity-temperature trends of the nanofluids. In fact, if the base fluid

involved was the highly viscous type such as glycerol, the addition of nanoparticles did not impair

the established relationship between the viscosities of nanofluids and temperature as shown in Fig.

5 [189].

29
Volume fraction

Volumetric concentration is the amount of nanoparticles dispersed in base fluid, usually less than

10% volume of the base fluid. Adequate research efforts have gone into discovering the effects of

volume fraction of nanoparticles on thermophysical behaviours of nanofluids. Most investigations

reported so far show that an increase in volume fraction of nanoparticles increases the viscosity of

nanofluids [190]. For example, Chevalier et al. [94] studied the rheological behaviours of SiO2-

ethanol nanofluids in microchannels and observed a constant Newtonian behaviour of the nanofluid

over the range of volume fraction studied (1.1%-7%) and shear rate values of 5 × 103-5 × 104 s-1.

Regarding the evolution of viscosity of the nanofluid with the said volume fraction range, there

was a direct relationship, i.e. as the volume fraction was increased, the viscosity also increased.

Corcione [111] analysed different experimental data from diverse literature for possible parameters

that affect the viscosity of nanofluids. His observation centred on the significant increase in the

viscosity as the volume fraction increases. Phuoc and Massoudi [110] observed experimentally that

for Fe2O3-deionised water, the nanofluid volume concentration of Fe2O3 is a critical parameter that

influences the viscosity of the nanofluid. Across the shear rate range (13.2-264 s-1) tested at room

temperature, there was a noticeable viscosity increase for 1%-4% volume fraction concentration.

Likewise, viscosity of Al2O3-PG at three different volume fractions was experimentally

investigated by Prasher et al. [191]. Their experimental data were comparable with data from Das

et al. [192] and Wang et al. [193], showing the strong dependence of viscosity enhancement on

volume concentration. It was further reported that at volume fraction of less than 4% investigated,

the Al2O3-PG nanofluid behaviour was Newtonian and viscosity increased with an increase in

volume fraction. Some of the most recent works also corroborate this finding [21, 194, 195], except

for a more recent experimental work by Suganthi et al. [96] on the viscosity of ZnO-PG nanofluid.

Contrary to most publications, the authors discovered that the addition of ZnO nanoparticles to PG

up to 2% volume fraction reduced the viscosity below the viscosity of the base fluid (PG). They

offered an explanation in line with the bonding characteristics created between the ZnO

30
nanoparticles and the PG molecules. The hydrogen-bonding network that existed between the PG

molecules was weakened by the introduction of ZnO nanoparticles, which translates to viscosity

reduction. The effect is similar to the influence of increased temperature on the intermolecular

bonding of nanofluids.

Generally, the observed increase in viscosity with volume fraction of nanoparticles could be

explained in view of the fact that an increase in volume of particles dispersed in base fluid will

result in a pronounced drag effect on individual particles due to Brownian motion. Therefore, the

overall drag effect present in the medium will be increased, which, in turn, will lead to an increase

in dissipation of energy and the consequence of this is the observed increase in the nanofluid

viscosity. Furthermore, it can be explained by exploring the particle surface charge mechanism in

relation to the suspension medium. When charged nanoparticles are dispersed into polar base fluid

for instance, the attraction of counterion onto the nanoparticle surface is likely to occur and this

process creates the formation of electrical double layer (EDL). An increase in nanoparticle

concentration will reduce the interparticle distance and by extension the distance between the

EDLs. Therefore, the force of interaction between the EDLs known as electroviscous force

introduces an additional increase in viscosity [132]. Other forces such as solvation, hydration,

hydrophilic and hydrophobic forces become very important and influence the rheology of

nanofluids when the interparticle distance is reduced due to an increase in volume fraction [196].

Particle agglomeration has also been argued as one of the causes of an increase in viscosity of

nanofluids. According to Chen et al. [197], a well-dispersed nanofluid suspension shows lower

viscosity compared with the corresponding agglomerated suspension. An increase of volume

concentration beyond the dilute regime heightens the tendency of agglomeration in nanofluid

systems, especially when the Van der Waals force of attraction is significant. When agglomeration

occurs, it forms a porous particle with the liquid of the suspending medium filling the interstices.

31
This immobile additional liquid in the interstices causes an increase in the effective volume

fraction, which leads to viscosity increase [132].

Shear rate

The control of the flow property of suspensions is very important in many industrial applications,

such as in the manufacturing of paint, crude oil drilling, crude and petroleum product

transportation, and in food and consumer products. Rheological studies of these products give

insights into the type of control that needs to be applied for efficient product transportation and

delivery. For instance, in the petroleum industry where hydrate formation in crude oil

transportation creates a flow problem or sometimes total blockage of the flow path [198], the

knowledge of the behaviour of hydrate slurry to different shear stress/shear rate among other

influencing parameters will make the delivery more efficient and less expensive [199].

The rheology of numerous nanofluids containing different nanoparticles including carbon

nanotubes has been studied over the past years [81, 84, 200]. Different rheological characteristics

have been reported ranging from Newtonian to non-Newtonian shear thinning, shear thickening,

thixotropic, etc. [81, 201]. In 2003, Tseng and Lin [83] investigated the rheology and structure of

TiO2-water nanofluid and found that within the volume fraction investigated (5-12%), the

nanofluid exhibited non-Newtonian viscoplastic fluid behaviour as presented in Fig. 6 (a). As

nanoparticles are added to the Newtonian base fluid up to 5%, an initial yield stress was needed to

be exceeded before flow could be achieved. This feature puts the nanofluids in the viscoplastic

non-Newtonian regime. When flow occurs, increase in the shearing rate from 0-1000 s-1 clearly

shows shear thinning structure except at 5% and shear rate ≥ 700 s-1 (Fig. 6 (b)). During this

shearing process, agglomerate structures formed in the nanofluid are broken down until they form

an ordered arrangement without agglomeration within the limits of high shearing rates [83]. To

buttress this point, similar trends and explanations have been offered in a more recent investigation

by Yang et al. [202], who investigated the rheology of diamond and Al2O3 nanofluids relative to

32
the effect of agglomeration. Two Newtonian base fluids were employed, namely silicone oil

(Syltherm 800) and DI-water. From their findings, the addition of as low as 0.35 and 0.24% of

diamond and Al2O3 nanoparticles without stabiliser turn the fluid to non-Newtonian with shear

thinning characteristics. Interestingly, in the stabilised Al2O3-DI-water nanofluids sample at 1.28%

volume fraction, the nanofluid behaved as a Newtonian fluid while the non-stabilised counterpart

clearly showed a shear thinning phenomenon without thixotropy. As stated earlier, the Newtonian

characteristic of the stabilised sample was due to the ordered structure of the nanoparticles in

suspension (i.e. devoid of agglomeration), while in the non-stabilised sample, there was

agglomeration due to dominance of the Van der Waals force of attraction and at low shear rate, it

showed high resistance to flow. As the shear rate increased, the agglomerates were broken down

and the immobile liquids within the interstices of the porous agglomerates were released, which

further reduced the viscosity alongside the ordered structure that was created at high shearing rate.

In another recent paper, Aladag et al. [108] showed that shearing time also influenced the internal

microstructure of nanofluids. Investigation on CNT- and Al2O3-water-based nanofluids at three

different shearing times and shear rates up to ~ 4000 s-1, showed that stabilised CNT and Al2O3 in

water responded with shear thinning, thixotropic, and shear thickening, thixotropic, phenomena

respectively. During the ramp up shear rate (low-high) there was deagglomeration and/or

realignment of agglomerated nanoparticles leading to the shear thinning behaviour. The

corresponding shear stress to the ramp down shear rate (high-low) was lower, which signifies the

thixotropic characteristics of the nanofluids as depicted in Fig. 7(a). But in the Al2O3 samples, the

characteristic was shear thickening with thixotropic phenomena as shown in Fig. 7(b). Based on the

shearing time experiments, it was clear from Fig. 7(b) that when the shearing time was relaxed,

sufficient time was provided for the rebuilding of the particle structure, which gave rise to the

increased ramp down shear stress shown for samples sheared for 180 and 240 seconds respectively.

Another work that showed shear thickening behaviour is the dispersion of Al2O3 in R141b

refrigerant up to 0.15% volume fraction [201].

33
Size of nanoparticles

A size of 1-100 nm particles is desired for nanofluid suspensions [203]. However, particles larger

than 100 nm have also been experimented with [82, 204]. Nguyen et al. [205] investigated the

effect of particle size (36 and 47 nm) on the viscosity of Al2O3-water nanofluids and reported that

the viscosity of both particles were similar at volume fraction below 4%. However, at higher

volume fraction, the viscosity was clearly higher in 47 nm. He et al. [206] also observed that a

bigger agglomerated size of TiO2 had higher viscosity compared with a smaller agglomerated size.

Contradictory reports [94, 207–209] have shown that smaller particle size led to an increase in

viscosity. The results from these reports appear to be reasonable given the following explanation.

When nanoparticles are dispersed in fluid medium, two major interactions are possible: particle-

fluid interaction and particle-particle interaction. These interactions have been termed first and

second electroviscous effects respectively [128]. Another prevalent interaction is the Van der

Waals force of attraction between particles. The electroviscous effect (EVE) present in the

nanofluid medium determines the agglomeration and hence the degree of Brownian motion effects

of the particles. For example, if the particle concentration is fixed, a reduced size of nanoparticles

translates into an increased overall surface area of solid-liquid interaction and also solid-solid

interaction, i.e. there will be an increase in the EVE present in the nanofluids, which, in turn, gives

rise to an increase in viscosity. Similarly, if the overall surface area of solid-liquid interface is

reduced, the EVE will be reduced. Thus, a reduction in viscosity had been observed experimentally

(Fig. 8) due to the increased size of nanoparticles [94, 132, 207–211]. A bigger particle size giving

a higher viscosity could be the result of particle agglomeration, as one of the reports clearly stated

[206].

In Eq. (1), the intrinsic viscosity [η] was provided as 2.5 for uncharged hard spherical particles.

However, due to the different processes of nanoparticle preparation [136, 156], the most correct

opinion is that nanoparticles always carry charges and in the absence or reduced strength of EVE,

34
there will be pH domination of the viscosity evolution in the nanofluid, which may further the

enhancement of cluster formation (i.e. agglomeration of nanoparticles). The shape of agglomerates

determines the intrinsic viscosity and it has been shown to be more than 2.5 when the agglomerate

is not spherical [212]. Anoop et al. [132] proposed that the intrinsic viscosity is not a shape

function, rather an EVE and agglomeration function. Eq. (45) was derived for the electroviscous

intrinsic viscosity and Eq. (46) for the agglomeration intrinsic viscosity:

 EV    1  p  (45)

 a     fa (46)

where p is the electroviscous coefficient and f a is the agglomeration factor, taken to be two

because the agglomerated nanoparticles doubled the starting size nanoparticles.

Shape of nanoparticles

Virtually all reported works on the thermophysical properties of nanofluids either assumed or

claimed their nanoparticles were spherical in shape except for those that worked with carbon

nanotubes (CNTs) and a few other works on different nanoparticles [98, 213]. Shape factor, which

is a measure of total surface area of particles, is related to the degree of solid-liquid interactions at

the interface. Timofeeva et al. [98] note that as the sphericity of particles reduces, there is a

corresponding increase in shape factor. Consequently, a larger surface area is presented for active

solid-liquid interactions. The same authors presented a correlation to describe the viscosity of

alumina-EG/water nanofluid with varying shapes. Although the proposed equation (Eq. 47) is more

of a replica of Batchelor’s equation (Eq. 48) [55], the authors elucidated that the coefficients A1

and A2 were higher than what was obtainable with Batchelor’s [55] even for spherical and non-

reacting particles, and these coefficients also varied with particle shapes.

nf  o 1  A1  A22  (47)

nf  o 1     6.5 2  (48)

35
The variation in viscosity as presented in [98] is the only data available with regard to different

shapes of nanoparticles in the same experimental set-up and conditions. Although some researches

on colloidal suspensions of shapes other than spheres had been carried out [213–215], they only

considered rod-like shapes. The increase in shape factor of particles has been thought to be one of

the causes of increase in nanofluid viscosity. However, the evolution of viscosity as presented by

Timofeeva et al. [98] was anomalous and without a distinctive relationship with shape/shape factor

of the nanoparticles.

Another factor that could cause a difference in the viscosity data of different nanoparticle shapes

could lie in the aerodynamics of these shapes in the suspended medium. A streamlined shape poses

less resistance to flow as compared with an irregular shape or sharp-edged shapes that resist flow

because it is difficult for such shapes to fall in ordered streamlines and as such they dissipate more

energy, which results in increased viscosity of the suspension. In view of the above, it is imperative

that more researches be focused in this direction.

pH and electrical conductivity of suspension

pH becomes important in the engineering of nanofluids given the prospect of data available,

emphasising the effect of zeta potential on the thermophysical properties of nanofluids. Zeta

potential defines the electrokinetic potential of the EDL and this influences the electrostatic

behaviour of nanofluids. The effect of pH on nanofluids’ zeta potential and thickness of EDL was

part of the investigation carried out by Rubio-Hernández et al. [212]. Manipulating the pH values

has been shown to bring about viscosity decrease [98]. However, manipulation should be carefully

carried out with respect to its effect on zeta potential, because an increase in the viscosity of

nanofluid with changes in pH values was also reported [211]. This is probably due to the fact that

the effect of pH manipulations on zeta potential values was not monitored and must have affected

the stability of the suspension. It has been shown that the Van der Waals force of attraction

dominates the interaction between nanoparticles when the pH of nanofluid is at the isoelectric point

36
(IEP). Because at IEP, the zeta potential is equal to zero and this hastens the agglomeration rate

[178].

The particle surface charge Q is directly related to the zeta potential as presented in Eq. (49) and

the critical zeta potential (ζcritical), which defines the critical surface charge, is also affected by

volume concentration [129]. The effect of volume loading is such that the value of zeta potential is

moved from below critical to above critical, i.e. ζ < ζcritical – ζ > ζcritical, consequently, there is a

reduction in the electrical conductivity or sometimes a reduction to plateau [129].

Q  4 r o a (49)

where  r is the relative permittivity of the medium,  o is the vacuum permittivity of the medium, a

is the radius of the particle, ζ is the zeta potential of the nanofluids. At this juncture, the following

conclusion can be drawn:

 pH might be insignificantly affected by an increase in volume concentration [98, 212],

how-ever, it greatly affects the ζ potential, which equally reduces the electrical conductivity

of nanofluids;

 pH modification reduces the viscosity of nanofluids when carefully carried out such that the

stability of the suspension is not compromised; and

 synergetic research is needed to be carried out on the zeta potential, pH and electrical

conductivity effect on the viscosity of nanofluids.

Base fluid properties

Nanofluid viscosity amplification is a multifactorial issue. Base fluid properties such as density,

thermal conductivity, viscosity and pH are definitely part of the factors that affect the overall

properties of nanofluids such as stability, thermal conductivity and viscosity. As the name “base

fluid” implies, it is the basis on which all enhancements are built. Many conventional heat transfer

fluids have been experimented with different nanoparticles, in different mixing ratios and very

common in the literature. However, researches claiming that base fluid properties influence the

37
viscosity enhancement have not presented any exhaustive experimental data probing the extent to

which the intrinsic properties of base fluids influence the viscosity enhancement of nanofluids.

Nonetheless, Syam Sundar et al. [188] investigated the viscosity of nanofluids synthesised from

magnetic Fe3O2 and three different EG-water mixtures as the base fluids, namely EG:water –

60:40, 40:60 and 20:80. Enhancement of approximately 300% was recorded for the sample

prepared from the 60:40 ratio, which was higher than that of other nanofluids. Wang et al. [193]

measured the viscosity of Al2O3 dispersed in water and EG. In their experiment, the viscosity of

Al2O3-water increased by 20-30% at 3% volume fraction and it was dependent on the dispersion

method, while for the Al2O3-EG sample at 3.5%, the viscosity increase was approximately 40%.

Recently, Yu et al. [106] dispersed aluminium nitride into EG and PG as base fluids and studied

their thermal conductivity and rheology. The rheological study shows that the enhancement in EG

is approximately 15% higher than the enhancement in PG.

These researches buttress the implication of base fluid in nanofluid viscosity. However, as

conventional heat transfer fluids are the bases for nanofluid synthesis, the same nanoparticles (i.e.

particles from the same metal or oxide of metal, equal volume fraction, equal average size, etc.)

have exhibited different behaviours in the same base fluid medium. Therefore, it is expedient to

probe into this area more deeply and possibly the unresolved cause of viscosity enhancement of

nanofluids will be clarified.

Magnetorheological nanofluids (MRNF) – a smart lubricant

Nanofluids are not restricted to function as a heat transfer working fluid alone as the majority of its

widely reported applications implied. For instance, its applicability has been studied in motor

vehicle engine radiators [8, 127], heat pipes [121], solar applications [123], boiling and

condensation systems [69, 192], circular tube heat exchangers [119], etc. In all cases cited, it has

served as heat transfer agent alone. There is a high prospect of nanofluids doubling as lubricant

cum heat transfer agent [216]. For example, it can be very useful in rotating shaft lubrication,

journal bearings, bearing seals or even at the ball and socket joint in motor vehicles. However, all

38
the above-mentioned applications come with high thermal energy generation, which will lower the

viscosity of any lubricant fluid, including nanofluids. Magnetic nanofluids, also called ferrofluids,

magnetic fluids and/or MRNF [217, 218], are a type of nanofluids in which the suspended

nanoparticles are ferromagnetic nanoparticles, ferrimagnetic nanoparticles and metallic

nanoparticles that are influenced by magnetic fields [219]. These fluids have been proved to be

tuneable to produce the desired thermophysical characteristics in the presence of an external

magnetic field [181, 220, 221]. Issues about its stability are of great concern owing to the fact that

magnetic nanoparticles (α- and -Fe2O3, Fe3O4, CoFe2O4, Co, Fe-C, MnFe2O4, MgFe2O3, Fe3N, Ni,

Fe, Fe-Co or Ni-Fe) are very heavy. Therefore, settling down or sedimentation due to gravitational

effect is a common drawback of MRNF nanofluids. Even with relatively viscous basefluids [43],

they display non-Newtonian characteristics, which can be attributed to the formation of

agglomerates. This is one way to measure the stability of the magnetic nanofluids [181, 219]. To

overcome the stability problem, different methods of synthesising magnetic nanoparticles have

been explored, some of which are wet-grinding [222], chemical co-precipitation [223–225] and

solvothermal [43] methods. Co-precipitation is the most favoured because it is rapid and the size of

magnetic nanoparticles can be controlled flexibly and efficiently by parametric variation of inputs

into the co-precipitation reaction [219]. In addition, the problem of non-compatibility of surfactant

used in the preparation of the particles with the desired carrier fluid is eliminated. Also, particles

derived from other methods, such as thermal decomposition of organometallic compounds have

been shown to lose their magnetisation due to poor crystallisation and/or oxidation[226].

Correspondingly, consensus on the size of magnetic nanoparticles is that it should be ≤ 10 nm and

with this, sedimentation can be overcome because sedimentation velocity is directly proportional to

the square of the particle diameter (i.e. V  d 2 ) [181]. However, the problem created by size ≤ 10

nm is the exhibition of magnetic saturation by individual (magnetic or metallic) particles of this

size. It means even in the absence of an external magnetic field, interactions between particles will

still cause instability in their respective carrier/base fluids with effects like zippering and

39
agglomeration. Therefore, stability is usually instituted in magnetic nanofluids and nanofluids

generally, using sterical surface coating of nanoparticles with the addition of surfactant (surface

modifier) or electrostatically charging (pH modification) nanoparticles in order to create a

repulsive interaction between particles, thereby ensuring stability [217].

A stable MRNF is of immense significance as a lubrication fluid at extremely high working

temperature. Presently, there are commercialised magnetorheological fluids (MRF) (fluids that

contain magnetic microparticles) that can operate in an environment with temperature up to 200 oC

[227]. However, these MRF fluids are disadvantaged just like the early colloids; they have an

abrasive nature in service and they settle down quickly, hence the urgency and importance of stably

and properly engineered MRNFs for lubrication application. At the switching on and off of the

magnetic field, external to the site of application of these fluids (MRNFs), a renewed lubrication

effect can be created and maintained by manipulating the strength and uniformity of the magnetic

field, thereby creating not only an efficient lubrication (in terms of volume and power

requirements) even at high temperatures, but also a durable system. Owing to this tuneable

characteristic of MRNFs, it has been called smart fluid.

NUMERICAL STUDIES

Algorithm-Based Numerical Studies

Employing the recent high computational power and the knowledge of algorithm, a new window of

opportunity has been opened to the nanofluid research community on the prediction models for

thermal properties of these highly revered heat transfer fluids. Rudyak and Krasnolutskii [228]

recently performed a computer simulation employing a standard molecular dynamics method to

model the interaction between nanoparticles and the carrier fluid molecules. Based on the

interaction potentials (basically a geometric function) of the particle-molecule, they simulated the

viscosity of lithium-argon and aluminium-argon nanofluids.

40
In early works by Huseyin and Muhammet [229], Hojjat et al. [230] and Papari et al. [231],

artificial neural network (ANN) had been applied for the prediction of thermal conductivity of

different nanofluids with good agreement with the values obtainable in the literature. Less than a

year after, Mehrabi et al. [232], using an FCM-based neuro-fuzzy inference system and genetic

algorithm-polynomial neural network alongside experimental data, developed two new models for

the prediction of thermal conductivity of Al2O3-water nanofluids.

The use of ANN is now gaining momentum in the prediction of the thermal properties of

nanofluids including the viscosity of different nanofluids. This is imperative because viscosity is a

known determinant parameter in the effective usage and design based on nanofluids. Yousefi et al.

[233], Bahiraei et al. [234] and Hajir and Yousefi [235] have now successfully applied the neural

network and genetic algorithm to the modelling of nanofluid viscosity for different nanofluids.

Mehrabi et al. [236] developed four different models for the prediction of nanofluid viscosities for

four different water-based nanofluids (Al2O3, CuO, TiO2 and SiO2) based on the FCM-based

adaptive neuro-fuzzy inference system. The results of the respective models were compared with

the experimental data points and some prominent correlations from the literature. The degree of

prediction of the FCM-ANFIS models was good. Atashrouz et al. [237], using data from the

literature, recently modelled the viscosity of nine nanofluids based on the group method of data

handling and polynomial neural network system (hybrid GMDH-PNN). The experimental data

used were on Al2O3-water, Al2O3-EG, Al2O3-PG, CuO-water, CuO-EG:water, SiO2-water, TiO2-

EG, and TiO2-water. Nine models were presented for individual nanofluids with average absolute

relative deviation of 2.14%. Employing the GMDH-PNN algorithm, a model can be regarded as a

set of neurons in which different pairs of the neurons in each layer are connected through a

polynomial of the second order, which then produces new neurons in the following layer. This

representation can be employed in modelling and predicting highly non-linear experimental data as

an inputs-outputs system.

41
It should be mentioned that this is probably the only algorithm-based work on the prediction of the

thermal conductivity and viscosity of nanofluids so far, therefore, researchers need to do more

work in this regard to further enrich nanofluid research.

Algorithm-Based Selection of the Viscosity of Nanofluids

There is the possibility of creating a viscosity model database from which a selection can be made.

To select nanofluid viscosity models which conform to a set of predetermined criteria, necessitates

the use of a generic algorithm. Based on the diversity of the models available and used in

predicting the viscosity of nanofluids, there is an emphasised need to select an appropriate model.

For the selection of a viscosity model that is likely going to be used for prediction, the criteria and

selection flow chart in Fig. 9 must be implemented.

The algorithm for the selection of appropriate models consists of defining arrays and sub-arrays,

which can be used to represent the attributes of each model. In Fig. 10, R, S up to V are viscosity

models with features and attributes given as R1, R2...R6, S1, S2...S6, and V1, V2,…V6

respectively. Additional information is held in the sub-array, represented as r1, s1 and v1

respectively. The system of arrays and sub-arrays defined for each model renders the following

features, among others, in a structured form for decision-making: the empirical or theoretical

nature of the model, its error margin, a mechanism for its formulation, its specificity (i.e.

applicability to a range of nanofluids) and the range of parameters used in defining the models.

Priority indices as shown in Fig. 10 represent the definition of selection criteria. The parameters are

allocated index numbers and 0 precedes 1 in terms of priority. After the implementation of search

and sort algorithm, more than one model may be presented as fitting the criteria set in their order of

priority. Models selected are tested for performance as depicted in the flow chart of Fig. 9.

The actual problem lies in selecting models with a broad range of parameters. The decision-based

selection algorithm can contain an infinite range of parameters in its search routine, and can be

42
recursively implemented. This ultimately can be used to build up a database of appropriate models

and filter off redundant models.

CONCLUSIONS AND FUTURE DIRECTIONS

Conclusions

In this review, an attempt was made to throw more light on the understanding of nanofluid

viscosity models and their evolution from the classical to numerical models using the findings

available in the literature. The pioneering work of Einstein was developed to predict the viscosity

of non-interacting and hard microsphere suspension in dilute regime. The majority of the

subsequent classical works on the viscosity of suspension were carried out to further extend

Einstein’s work to a concentrated regime [55, 62, 68]. These efforts show the importance of the

influence of increase in volume fraction on the viscosity of suspension. However, none of the

classical models was developed for the estimation of nanofluid viscosity. Recently, a few new

theoretical models have been developed taking into consideration some of the characteristics of

nanofluids such as nanoparticle size, hydrodynamic volume fraction, particle density, aggregate

diameter and capping layer thickness. The empirical models were applied to characterise the

behaviour of nanofluids at different nanoparticle volume fractions, nanoparticles sizes,

temperatures and shear rates. It was noted that most empirical models were developed considering

a single parameter, mostly volume fraction or temperature. However, nanofluid viscosity is

affected by multiple factors such as volume fraction, particle size, particle shape, shearing rate,

shearing time, particle agglomeration, base fluid properties and nanolayer. Therefore, the review

further looked into the factors that affect nanofluid viscosity.

Generally, experimental investigations have shown that the addition of nanoparticles to

conventional heat transfer fluid leads to an increase in viscosity of the fluid. A further increase in

particle volume fraction has been shown to lead to non-linear increment in the viscosity (Fig. 2 and

3). The viscosity of nanofluids decreases with an increase in temperature and it is usually described

43
by an exponential curve (Fig. 5). It is desirable to prepare the nanofluid with particle size ≤ 100

nm, because a bigger size in the microregime hinders practical application due to abrasion, higher

rate of settlement, pressure drop in flow line and clogging of equipment. Experiments have shown

that nanoparticle size is another very important factor that influences the viscosity of nanofluids.

Typically, as the particle size reduces, the viscosity of suspension increases [94, 207–209]. The

increase has been ascribed to an increase in particle-particle interactions due to increased Brownian

motion and if the system is not sterically or electrostatically stabilised, there may be agglomeration,

which increases the viscosity as many studies have shown. Other influencing factors reviewed were

nanoparticle shape, shearing rate, pH and electrical conductivity, and base fluid.

The stability of nanofluids is essential for its eventual implementation for practical purposes.

Experimental results available have shown that the methods of nanofluid preparation especially

from the two-step method go a long way in dictating the stability of nanofluid. Fedele et al. [157]

showed that the ultrasonication method of nanoparticles dispersion was more effective than ball

milling when they prepared DI-water-based nanofluids of CuO, TiO2 and SWCNHs. The addition

of surfactant (dispersant or surface modifier) or pH modification was also shown to be effective

[157, 180]. However, there have been reports that surfactant and pH sometimes lead to an

unnecessary increase in viscosity of nanofluids when not properly modified [82, 211]. This review

also showed that the present methods of investigating the stability of nanofluids are deficient

because they do not represent what is obtainable in the practical situations that nanofluids might be

subjected to.

Lastly, the emerging numerical methods available for the prediction of nanofluids viscosity were

also presented. Molecular dynamics simulation and the use of artificial intelligence (artificial

neural networks) are fast gaining ground due to their ability to model the non-linear nature of

viscosity data. The ANN-based models have shown good predictions with reduced % AARD and

44
can take a wide range of influencing factors into consideration compared with most empirical

models that are built around a single parameter such as temperature or volume fraction.

Regarding the composite of nanofluids, which can consist of different fluid bases and different

nanoparticles, different accurate correlations for different nanofluids for a specific range of volume

fraction, nanoparticle size and temperature (with or without surfactant) need to be developed.

Future direction

The review of Wong and De Leon [238] on the current and future trends of nanofluids dealt

extensively with the present potentials obtainable from the use of nanofluids. Others that have

reviewed the current areas of applications of nanofluids include Yu and Xie [239] and Wang and

Mujumdar [6]. The following are the key areas common to all the above-mentioned reviews:

– application of nanofluids in industrial cooling such as in heat exchangers, including nuclear

reactor cooling;

– use of nanofluid as motor vehicle coolants such as in engine oil, transmission fluid, grease

and radiator coolant;

– application of nanofluids in electronics cooling;

– nanofluids as fuel enhancer to boost energy recovery and reduce greenhouse and poisonous

gas emissions; application in space and defence industry; and

– use of nanofluid for biomedical applications such as nanodrug delivery, antibacterial

activities and in targeted cancer therapeutics.

All the above bullet points have been appropriately dealt with in more elaborate manner [238–240].

However, there are still few areas that need attention. Firstly, the enhanced properties of

nanofluids can be employed in boosting energy storage and recovery in solar ponds. A solar pond

consists of a body of fluid used in the collection and storage of thermal energy. Research has

shown the possibility of using saturated saltwater for heat and energy generation [241]. With the

invention of nanofluid, it is believed that heat transfer mechanism of a saltwater solar pond vis-à-

45
vis absorption and storage can be enhanced with the use of nanofluid. Recently, Al-Nimr and Al-

Dafaie [242] numerically investigated the use of Ag-water nanofluid and mineral oil in a two-layer

nanofluid solar pond. They found that high absorption coefficient is required for better

performance and this was observed to be directly related to an increase in volume fraction.

Therefore, at high nanoparticle concentration, the nanofluid solar pond was found to be more

effective that the conventional saltwater solar pond. In this regard, there is the need for

experimental validation and more robust numerical research to further unlock nanofluid potential

for use in solar ponds. Secondly, nanofluid stability is an important issue as already highlighted in

this review. The effect of pH on stability cannot be underestimated because through pH

modification, nanofluid suspension can be electrostatically stabilised [179, 180]. Studies have

shown that as the pH of nanofluids moves further away from the IEP, the higher the absolute value

of zeta potential, which is the measure of stability. It should be noted that most of these studies

were carried out at room temperature. Until very recently, there has been no study on the effect of

temperature on the pH of nanofluids [189, 243]. These studies [189, 243] have shown that

temperature has a considerable effect on the pH of nanofluids. Therefore, it is expedient to

intensify research efforts in this regard with the hope of producing nanofluid at a pH and

temperature where zeta potential is in the stable region, such that if the nanofluid is kept at this

temperature for a period, its stability can be studied. Lastly, the need to study sonication energy

and nanofluid stability with respect to viscosity enhancement is of paramount importance. In many

experiments, as shown in this review, researchers just chose an arbitrary number of hours for their

ultrasonication. This method is not standard as is the reporting as well. Further experimental

investigation should be carried out to study the impact of ultrasonication on the viscosity and

stability of nanofluids. Above all, the experimental values on ultrasonication should be reported on

energy/volume (energy density) of nanofluids prepared to enhance repeatability of experiment by

other researchers.

46
ACKNOWLEDGEMENTS

The funding obtained from the National Research Foundation of South Africa (NRF), Stellenbosch

University/University of Pretoria Solar Hub, CSIR, EEDSM Hub, NAC and IRT SEED is

acknowledged and duly appreciated.

NOMENCLATURE

a particle radius

aa effective radius of aggregate

ai empirical constants

A empirical parameter

bI characteristics of electrolyte

B empirical parameter

C1 correction factor

C empirical constant

CL Chen and Law model

CNT carbon nanotube

dp nanoparticle diameter

df base fluid molecular diameter

di diameter of the ith particle size

Da aggregate diameter

Dai ith aggregate diameter

DE dielectric constant

Dx average particle diameter in

DI deionized

e charge

E empirical constant

47
Ea activation energy

EDL electrical double layer

EVE electroviscous effect

fa agglomeration factor

h thickness of nanolayer

hs minimum separation distance between two spheres

HPH high pressure homogenization

k specific conductivity

kb Stephan-Boltzmann constant

kH Hugging’s coefficient

m system property constant

MRF magnetorheological fluid

MRNF magnetorheological nanofluid

MWCNT multi-wall carbon nanotube

Ni number of particles with ith size diameter

p electroviscous coefficient

p average aspect ratio

PSD particle size distribution

Pi packing fraction of each class size i

Q particle surface charge

r capping layer thickness

R adjustable parameter,

Rg universal gas constant

SWCNH single-wall carbon nanohorn

SWCNT single-wall carbon nanotube

48
to initial time after preparation

t3 final time at phase separation

T suspension temperature

To reference temperature

vj binary packing coefficient

V sedimentation velocity

VB Brownian velocity

z number of different class of particle size in suspension

Greek Symbols

 empirical constant

 diffusion coefficient

 empirical constant

 shear rate

 ij binary packing fraction

 particle centre-centre distance

o permittivity of the vacuum

r relative permittivity of the medium

 intrinsic viscosity

 zeta potential

r relative viscosity

 temperature in degree Celsius

 empirical constant

e electrophoretic mobility

eff effective viscosity

49
i viscosity corresponding to ith class particle size distribution

nf nanofluid viscosity

o suspending medium viscosity

s specific viscosity

 intrinsic viscosity at infinite shear rate

,T viscosity at infinite temperature

 viscosity of fluid droplet

1   4 dimensionless parameters

f base fluid density

p nanoparticle density

 empirical constant

 shear stress

o yield stress

 crowding factor

 packing geometry of inorganic materials

 particle volume fraction

a aggregate volume fraction

eff effective volume fraction

h hydrodynamic volume fraction

i volume fraction corresponding to ith class particle size distribution

m maximum particle volume fraction

ma packing fraction of aggregates

50
z ultimate packing fraction

 nanoparticle mass fraction

i mass fraction of the aggregate i

 empirical constant

 empirical constant

Subscripts

a aggregate

b Boltzmann

B Brownian

eff effective

EV electroviscous

f fluid

h hydrodynamic

i ith class

j jth class

m maximum

nf nanofluid

o reference

p particle

r relative

s separation

Superscripts

mono monomodal particle distribution

n empirical constant

x particle size distribution average

51
z number of modal suspensions (mono, bi or multi)

REFERENCES

[1] Maxwell, J.C., Electricity and Magnetism, Oxford, Clarendon, 1873.

[2] Masoumi, N., Sohrabi, N., and Behzadmehr, A., A new model for calculating the effective

viscosity of nanofluids, Journal of Physics D: Applied Physics, vol. 42, no. 5, pp. 0555011–

0555016, 2009.

[3] Choi, S.U.S., Enhancing thermal conductivity of fluid with nanoparticles, ASME. pp. 99–

109. , New York, 1995.

[4] Lee, S., Choi, U.S., and Li, S., Measuring thermal conductivity of fluids containing oxide

nanoparticles, Journal of Heat Transfer, vol. 121, pp. 280–289, 1999.

[5] Murshed, S.M.S., Leong, K.C., and Yang, C., Thermophysical and electrokinetci properties

of nanofluids - A critical review, Applied Thermal Engineering, vol. 28, pp. 2109–2125,

2008.

[6] Xiang-Qi, W., and Arun, S.M., A review on nanofluids - part II: experiments and

applications, Brazilian Journal of Chemical Engineering, vol. 24, no. 04, pp. 631–648,

2008.

[7] Vajjha, R.S., Das, D.K., and Namburu, P.K., Numerical study of fluid dynamic and heat

transfer performance of Al2O3 and CuO nanofluids in the flat tubes of a radiator,

International Journal of Heat and Fluid Flow, vol. 31, no. 4, pp. 613–621, 2010.

[8] Peyghambarzadeh, S.M., Hashemabadi, S.H., Jamnani, M.S., and Hoseini, S.M., Improving

the cooling performance of automobile radiator with Al2O3/water nanofluid, Applied

Thermal Engineering, vol. 31, no. 10, pp. 1833–1838, 2011.

[9] Nguyen, C.T., Roy, G., Gauthier, C., and Galanis, N., Heat transfer enhancement using

Al2O3-water nanofluid for an electronic liquid cooling system, Applied Thermal

Engineering, vol. 27, pp. 1501–1506, 2007.

52
[10] Tsai, C.Y., Chien, H.T., Ding, P.P., Chan, B., Luh, T.Y., and Chen, P.H., Effect of structural

character of gold nanoparticles in nanofluid on heat pipe thermal perfomance, Material

Letters, vol. 58, no. 9, pp. 1461–1465, 2004.

[11] Ma, H.B., Wilson, C., Borgmeyer, B., Park, K., and Yu, Q., Effect of nanofluid on the heat

transport capability in an oscilating heat pipe, Applied Physics Letters, vol. 88, no. 143116,

pp. 1–3, 2006.

[12] Nguyen, C.T., Roy, G., Lajoie, P.R., and Maiga, S.E.B., Nanofluids Heat Transfer

Performance for Cooling of High Heat Output Microprocessor, Proc. of 3rd

IASME/WSEAS Int. Conf. on Heat Transfer, Thermal Engineering and Environment. pp.

160–165. , Corfu, Greece, 2005.

[13] Jain, P., El-Sayed, I., and El-Sayed, M., Au nanoparticles target cancer, nano today, vol. 2,

no. 1, pp. 18–29, 2007.

[14] Sun, X., Liu, Z., Welsher, K., Robinson, J.T., Goodwin, A., Zaric, S., and Dai, H., Nano-

Graphene Oxide for Cellular Imaging and Drug Delivery., Nano research, vol. 1, no. 3, pp.

203–212, 2008.

[15] Pankhurst, Q. a, Connolly, J., Jones, S.K., and Dobson, J., Applications of magnetic

nanoparticles in biomedicine, Journal of Physics D: Applied Physics, vol. 36, no. 13, pp.

R167–R181, 2003.

[16] Gupta, A.K., and Gupta, M., Synthesis and surface engineering of iron oxide nanoparticles

for biomedical applications., Biomaterials, vol. 26, no. 18, pp. 3995–4021, 2005.

[17] Lee, J.-H., Hwang, K.S., Jang, S.P., Lee, B.H., Kim, J.H., Choi, S.U.S.S., and Choi, C.J.,

Effective viscosities and thermal conductivities of aqueous nanofluids containing low

volume concentrations of Al2O3 nanoparticles, International Journal of Heat and Mass

Transfer, vol. 51, no. 11-12, pp. 2651–2656, 2008.

53
[18] Yu, W., Xie, H., Chen, L., and Li, Y., Investigation of thermal conductivity and viscosity of

ethylene glycol based ZnO nanofluid, Thermochimica Acta, vol. 491, no. 1-2, pp. 92–96,

2009.

[19] Eastman, J. a. A., Phillpot, S.R.R., Choi, S.U.S.U.S., and Keblinski, P., Thermal Transport

in Nanofluids, Annual Review of Materials Research, vol. 34, no. 1, pp. 219–246, 2004.

[20] Maïga, S.E., Palm, S.J., Nguyen, C.T., Roy, G., and Galanis, N., Heat transfer enhancement

by using nanofluids in forced convection flows, International Journal fo Heat and Fluid

Flow, vol. 26, pp. 530–546, 2005.

[21] Heyhat, M.M.M., Kowsary, F., Rashidi, A.M.M., Memenpour, M.H., Amrollahi, A., and

Momenpour, M.H., Exeprimental investigation of laminar convective heat transfer and

pressure drop of water-based Al2O3 nanofluids in fuly developed flow regime, Experimental

Thermal and Fluid Science, vol. 44, pp. 483–489, 2013.

[22] Quaresma, J.N.N., Macêdo, E.N., da Fonseca, H.M., Orlande, H.R.B., Cotta, R.M., Macedo,

E.N., and Fonseca, H.M.D., An Analysis of Heat Conduction Models for Nanofluids, Heat

Transfer Engineering, vol. 31, no. 14, pp. 1125–1136, 2010.

[23] Wang, J.J., Zheng, R.T., Gao, J.W., and Chen, G., Heat conduction mechanisms in

nanofluids and suspensions, Nano Today, vol. 7, no. 2, pp. 124–136, 2012.

[24] Wen, D., Lin, G., Vafaei, S., and Zhang, K., Review of nanofluids for heat transfer

applications, Particuology, vol. 7, no. 2, pp. 141–150, 2009.

[25] Philip, J., and Shima, P.D., Thermal properties of nanofluids, Advances in Colloid and

Interface Science, vol. 183, pp. 30–45, 2012.

[26] Buschmann, M.H., Thermal conductivity and heat transfer of ceramic nanofluids,

International Journal of Thermal Sciences, vol. 62, no. 0, pp. 19–28, 2012.

[27] Mallick, S.S., Mishra, A., and Kundan, L., An investigation into modelling thermal

conductivity for alumina-water nanofluids, Powder Technology, vol. 233, pp. 234–244,

2013.

54
[28] Fan, J., and Wang, L., Heat conduction in nanofluids: Structure-property correlation,

International Journal of Heat and Mass Transfer, vol. 54, no. 19, pp. 4349–4359, 2011.

[29] Saleh, R., Putra, N., Prakoso, S.P., and Septiadi, W.N., Experimental investigation of

thermal conductivity and heat pipe thermal performance of ZnO nanofluids, International

Journal of Thermal Sciences, vol. 63, no. 0, pp. 125–132, 2013.

[30] Altan, C.L., Elkatmis, A., Yuksel, M., Aslan, N., and Bucak, S., Enhancement of thermal

conductivity upon application of magnetic field to Fe3O4 nanofluids, Journal of Applied

Physics, vol. 110, no. 093917, pp. 1–8, 2011.

[31] Gupta, S.S., Siva, M. V, Krishnan, S., Sreeprasad, T.S., Singh, P.K., Pradeep, T., and Das,

S.K., Thermal conductivity enhancement of nanofluids containing graphene nanosheets,

Journal of Heat Transfer, vol. 110, no. 084302, pp. 1–7, 2011.

[32] Chandrasekar, M., Suresh, S., and Bose, A.C., Experimental investigations and theoretical

determination of thermal conductivity and viscosity of Al2O3/water nanofluid, Experimental

Thermal and Fluid Science, vol. 34, no. 2, pp. 210–216, 2010.

[33] Hosseini, S., Moghadassi, A., and Henneke, D.E., A new dimensionless group model for

determining the viscosity of nanofluids, J. Therm Anal calorim, vol. 100, pp. 873–877,

2010.

[34] Namburu, P.K., Kulkarni, D.P., Misra, D., and Das, D.K., Viscosity of copper oxide

nanoparticles dispersed in ethylene glycol and water mixture, Experimental Thermal and

Fluid Science, vol. 32, pp. 397–402, 2007.

[35] Rashin, M.N., and Hemalatha, J., Viscosity studies on novel copper oxide – coconut oil

nanofluid, Experimental Thermal and Fluid Science, vol. 48, pp. 67–72, 2013.

[36] Mahbubul, I.M., Saidur, R., and Amalina, M.A., Latest developments on the viscosity of

nanofluids, International Journal of Heat and Mass Transfer, vol. 55, no. 4, pp. 874–885,

2012.

55
[37] Sundar, L.S., Sharma, K.V., Naik, M.T., and Singh, M.K., Empirical and theoretical

correlations on viscosity of nanofluids: A review, Renewable and Sustainable Energy

Reviews, vol. 25, pp. 670–686, 2013.

[38] Das, S.K., Choi, S.U.S., and Patel, H.E., Heat transfer in nanofluids - A review, Heat

Transfer Engineering, vol. 27, no. 10, pp. 3–19, 2006.

[39] Okhio, C., Hodges, D., and Black, J., Review of Literature on Nanofluid Flow and Heat

Transfer Properties, Cyber Journals: Multidisciplinary Journals in Science and Technology,

JOurnal of Selected Areas in Nanotechnology (JSAN), pp. 1–8, 2010.

[40] Wang, X.-Q., and Mujumdar, A.S., Heat transfer characteristics of nanofluids: a review,

International Journal of Thermal Sciences, vol. 46, no. 1, pp. 1–19, 2007.

[41] Wang, X., and Mujumdar, A., A review on nanofluids-part I: theoretical and numerical

investigations, Brazilian Journal of Chemical Engineering, vol. 25, no. 04, pp. 613–630,

2008.

[42] Moosavi, M., Goharshadi, E.K., and Youssefi, A., Fabrication, characterization and

measurement of some physicochemical properties of ZnO nanofluids, International journal

of Heat and Fluid Flow, vol. 31, no. 4, pp. 599–605, 2010.

[43] Abareshi, M., Sajjadi, S.H., Zebarjad, S.M., and Goharshadi, E.K., Fabrication,

characterization, and measurement of viscosity of α-Fe2O3-glycerol nanofluids, Journal of

Molecular Liquids, vol. 163, no. 1, pp. 27–32, 2011.

[44] Kole, M., and Dey, T.K., Effect of aggregation on the viscosity of copper oxide-gear oil

nanofluids, International Journal of Thermal Sciences, vol. 50, no. 9, pp. 1741–1747, 2011.

[45] Avsec, J., and Oblak, M., The calculation of thermal conductivity, viscosity and

thermodynamic properties for nanofluids on the basis of statistical nanomechanics,

International Journal of Heat and Mass Transfer, vol. 50, no. 21-22, pp. 4331–4341, 2007.

56
[46] Meyer, J.P., Nwosu, P., Sharifpur, M., and Ntumba, T., Parametric analysis of effective

viscosity models for nanofluids, Proceedings of the ASME 2012 International Mechanical

Engineering Congress & Exposition. pp. 1–9. , Houston, TX, 2012.

[47] Einstein, A., A New Determination of Molecular Dimensions, Annalen der Physik, vol. 4,

no. 19, pp. 37–62, 1906.

[48] Smoluchowski, M. V, Theoretishche Bemerkungen uber die Viskositat der Kolloide,

Kolloidzschr, vol. 80, pp. 190–195, 1916.

[49] Jeffery, G.B., The Motion of Ellipsoidal Particles Immersed in a Viscous Fluid, Proceedings

of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol. 102, no. 715,

pp. 161–179, 1922.

[50] Taylor, G.I., The viscosity of a fluid containign small drops of another fluid, Proc. Roy.

Soc., vol. 138, no. 834, pp. 41–48, 1932.

[51] Brinkman, H.C., The viscosity of concentrated suspensions and solutions, Journal of

Chemical Physics, vol. 20, no. 4, pp. 571, 1952.

[52] Vand, V., Viscosity of solutions and suspensions. I. Theory, The Journal of Physical

Chemistry, pp. 277–299, 1948.

[53] Mooney, M., The viscosity of a concentrated suspension of spherical particles, Journal of

Colloid Science, no. 113, pp. 3–4, 1951.

[54] Roscoe, R., The viscosity of suspensions of rigid spheres, Journal of Applied Physics, vol.

267, pp. 3–6, 1952.

[55] Batchelor, G., The effect of Brownian motion on the bulk stress in the suspension of

sphrerical particles, Journal of Fluid Mechanics, vol. 83, no. 01, pp. 97–117, 1977.

[56] Krieger, I., and Dougherty, T., A mechanicsm for non-Newtonian flow in suspensions of

rigid spheres, Trans. Soc. Rheology, vol. 3, pp. 137–152, 1959.

[57] Lundgren, T.S., Slow flow through stationary random beds and suspensions of spheres,

Journal of Fluid Mechanics, vol. 51, no. 02, pp. 273–299, 1972.

57
[58] Graham, A.L., On the viscosity of suspensions of solid spheres, Applied Scientific Research,

vol. 37, no. june, pp. 275–286, 1981.

[59] Saitô, N., Concentration dependence of the viscosity of high polymer solution, J. Phys. Soc.

Jpn, vol. 5, pp. 4–8, 1950.

[60] Hatschek, E., The general theory of viscosity of two phase systems, Trans. Faraday Soc.,

vol. 9, no. 2, pp. 80–92, 1913.

[61] Thomas, C.U., and Muthukumar, M., Three-body hydrodynamic effects on viscosity of

suspensions of spheres, The Journal of Chemical Physics, vol. 94, no. 7, pp. 5180, 1991.

[62] Frankel, N., and Acrivos, A., On the viscosity of a concentrated suspension of solid spheres,

Chemical Engineering Science, vol. 22, pp. 847–853, 1967.

[63] Booth, F., The electroviscous effect for suspensions of solid spherical particles, Proceedings

of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol. 203, no.

1075, pp. 533–551, 1950.

[64] Bull, H., The electroviscous effect in egg albumin solutions, Trans. Faraday Soc., no. 80,

pp. 80–84, 1940.

[65] Ward, S., and Whitmore, R., Studies of the viscosity and sedimentation of suspensions Part

1.-The viscosity of suspension of spherical particles, British Journal of Applied Physics, vol.

286, pp. 1–6, 1950.

[66] Williams, P.S., Flow of concentrated suspensions, Journal of Applied Chemistry, vol. 3, no.

3, pp. 120–128, 1953.

[67] Maron, S., and Fok, S., Effect of concentration on flow behavior of glass sphere

suspensions, Journal of Colloid Science, pp. 540–542, 1953.

[68] Chen, H., Ding, Y., and Tan, C., Rheological behaviour of nanofluids, New Journal of

Physics, vol. 9, no. 10, pp. 367–391, 2007.

[69] Kole, M., and Dey, T.K., Thermophysical and pool boiling characteristics of ZnO-ethylene

glycol nanofluids, International Journal of Thermal Sciences, vol. 62, pp. 61–70, 2012.

58
[70] Farris, R., Prediction of the viscosity of multimodal suspension from unimodal viscosity

data, Trans. Soc. Rheol, vol. 12, pp. 281–301, 1968.

[71] Chong, J., Christiansen, E., and Baer, A., Rheology of concentrated suspensions, Journal of

applied polymer science, vol. 15, no. 1971, pp. 2007–2021, 1971.

[72] Storms, R.F., Ramarao, B.V., and Weiland, R.H., Low shear rate viscosity of bimodally

dispersed suspensions, Powder Technology, vol. 63, no. 3, pp. 247–259, 1990.

[73] Dames, B., Morrison, B.R., and Willenbacher, N., An empirical model predicting the

viscosity of highly concentrated, bimodal dispersions with colloidal interactions, Rheologica

Acta, vol. 40, no. 5, pp. 434–440, 2001.

[74] Muralidharan, G., and Runkana, V., Rheological Modeling of Spherical Polymeric Gels and

Dispersions Incorporating the Influence of Particle Size Distribution and Surface Forces,

Industrial & Engineering Chemistry Research, vol. 48, no. 19, pp. 8805–8811, 2009.

[75] Servais, C., Jones, R., and Roberts, I., The influence of particle size distribution on the

processing of food, Journal of Food Engineering, vol. 51, pp. 201–208, 2002.

[76] Kole, M., and Dey, T.K., Viscosity of alumina nanoparticles dispersed in car engine coolant,

Experimental Thermal and Fluid Science, vol. 34, no. 6, pp. 677–683, 2010.

[77] Prasher, R., Bhattacharya, P., and Phelan, P.E., Brownian-Motion-Based Convective-

Conductive Model for the Effective Thermal Conductivity of Nanofluids, Journal of Heat

Transfer, vol. 128, no. 6, pp. 588–595, 2006.

[78] Graf, W.H., Hydraulics of Sediment Transport, McGraw-Hill, New York, 1971.

[79] Batchelor, G.K., and Green, J.T., The hydrodynamic interaction of two small freely-moving

spheres in a linear flow field, Journal of Fluid Mechanics, vol. 56, no. 3, pp. 401–427, 1972.

[80] Cheng, N., and Law, A., Exponential formula for computing effective viscosity, Powder

technology, vol. 129, pp. 156–160, 2003.

[81] Chen, H., Ding, Y., He, Y., and Tan, C., Rheological behaviour of ethylene glycol based

titania nanofluids, Chemical Physics Letters, vol. 444, pp. 333–337, 2007.

59
[82] Tseng, W.J., and Chen, C., Effect of polymeric dispersant on rheological behavior of nickel-

terpineol suspensions, Materials Science and Engineering: A, vol. 347, no. 1, pp. 145–153,

2003.

[83] Tseng, W.J., and Lin, K., Rheology and colloidal structure of aqueous TiO2 nanoparticle

suspensions, Materials Science and Engineering: A, vol. 355, pp. 186–192, 2003.

[84] Chen, H., Ding, Y., Lapkin, A., and Fan, X., Rheological behaviour of ethylene glycol-

titanate nanotube nanofluids, Journal of Nanoparticle Research, vol. 11, pp. 1513–1520,

2009.

[85] Horri, B.A., Ranganathan, P., Selomulya, C., and Wang, H., A new empirical viscosity

model for ceramic suspensions, Chemical Engineering Science, vol. 66, no. 12, pp. 2798–

2806, 2011.

[86] Sahoo, B.C., Vajjha, R.S., Ganguli, R., Chukwu, G.A., and Das, D.K., Determination of

Rheological Behaviour of Aluminium Oxide Nanofluid and Development of New Viscosity

Correlations, Petroleum Science and Technology, vol. 27, pp. 1757–1770, 2009.

[87] Nguyen, C.T., Desgranges, F., Galanis, N., Roy, G., Maré, T., Boucher, S., Angue Mintsa,

H., Maré, T., and Mintsa, H.A., Viscosity data for Al2O3 - water nanofluid - hysteresis: is

heat transfer enhancement using nanofluids reliable?, International Journal of Thermal

Sciences, vol. 47, no. 2, pp. 103–111, 2008.

[88] Hernández Battez, A., Viesca, J.L., González, R., García, A., Reddyhoff, T., and Higuera-

Garrido, a., Effect of Shear Rate, Temperature, and Particle Concentration on the

Rheological Properties of ZnO and ZrO2 Nanofluids, Tribology Transactions, vol. 57, no. 3,

pp. 489–495, 2014.

[89] Vakili-Nezhaad, G., and Dorany, A., Effect of Single-Walled Carbon Nanotube on the

Viscosity of Lubricants, Energy Procedia, vol. 14, no. 1998, pp. 512–517, 2012.

[90] Garg, J., Poudel, B., Chiesa, M., Gordon, J.B., Ma, J.J., Wang, J.B., Ren, Z.F., Kang, Y.T.,

Ohtani, H., Nanda, J., McKinley, G.H., and Chen, G., Enhanced thermal conductivity and

60
 viscosity of copper nanoparticles in ethylene glycol nanofluid, Journal of Applied Physics,

vol. 103, no. 7, pp. 074301, 2008.

[91] Godson, L., Raja, B., Lal, D.M., and Wongwises, S., Experimental Investigation on the

Thermal Conductivity and Viscosity of Silver-Deionized Water Nanofluid, Experimental

Heat Transfer, vol. 23, no. 4, pp. 317–332, 2010.

[92] Duangthongsuk, W., and Wongwises, S., Measurement of temperature-dependent thermal

conductivity and viscosity of TiO2-water nanofluids, Experimental Thermal and Fluid

Science, vol. 33, no. 4, pp. 706–714, 2009.

[93] Colla, L., Fedele, L., Scattolini, M., and Bobbo, S., Water-Based Fe2O3 Nanofluid

Characterization: Thermal Conductivity and Viscosity Measurements and Correlation,

Advances in Mechanical Engineering, vol. 2012, pp. 1–8, 2012.

[94] Chevalier, J., Tillement, O., and Ayela, F., Rheological properties of nanofluids flowing

through microchannels, Applied Physics Letters, vol. 91, no. 233103, pp. 1–3, 2007.

[95] Suganthi, K.S., and Rajan, K.S., Temperature induced changes in ZnO–water nanofluid:

Zeta potential, size distribution and viscosity profiles, International Journal of Heat and

Mass Transfer, vol. 55, no. 25-26, pp. 7969–7980, 2012.

[96] Suganthi, K., Anusha, N., and Rajan, K., Low viscous ZnO–propylene glycol nanofluid: a

potential coolant candidate, Journal of nanoparticle research, vol. 15, no. 1986, pp. 1986–

1–6, 2013.

[97] Singh, D., Timofeeva, E., Yu, W., Routbort, J., France, D., Smith, D., and Lopez-Cepero,

J.M., An investigation of silicon carbide-water nanofluid for heat transfer applications,

Journal of Applied Physics, vol. 105, no. 6, pp. 064306, 2009.

[98] Timofeeva, E. V., Routbort, J.L., and Singh, D., Particle shape effects on thermophysical

properties of alumina nanofluids, Journal of Applied Physics, vol. 106, no. 1, pp. 014304–1

– 10, 2009.

61
[99] Takeuchi, A., Kato, H., and Inoue, A., Vogel--Fulcher--Tammann plot for viscosity scaled

with temperature interval between actual and ideal glass transitions for metallic glasses in

liquid and supercooled liquid states, Intermetallics, vol. 18, pp. 406–411, 2010.

[100] Żyła, G., Witek, A., and Cholewa, M., Viscosity of diethylene glycol-based Y2O3

nanofluids, Journal of Experimental Nanoscience, pp. 1–8, 2013.

[101] Kulkarni, D., Das, D., and Chukwu, G.A., Temperature dependent rheological property of

copper oxide nanoparticles suspension (nanofluid), J. Nanosci. Nanotechnol., vol. 6, no. 4,

pp. 679–690, 2009.

[102] Yiamsawas, T., Dalkilic, A.S., Mahian, O., and Wongwises, S., Measurement and

Correlation of the Viscosity of Water-Based Al2O3 and TiO2 Nanofluids in High

Temperatures and Comparisons with Literature Reports, Journal of Dispersion Science and

Technology, vol. 34, no. 12, pp. 1697–1703, 2013.

[103] Hemmat Esfe, M., and Saedodin, S., An experimental investigation and new correlation of

viscosity of ZnO–EG nanofluid at various temperatures and different solid volume fractions,

Experimental Thermal and Fluid Science, vol. 55, pp. 1–5, 2014.

[104] Azmi, W.H., Sharma, K. V, Mamat, R., Alias, a B.S., and Misnon, I.I., Correlations for

thermal conductivity and viscosity of water based nanofluids, IOP Conference Series:

Materials Science and Engineering, vol. 36, pp. 012029, 2012.

[105] Khanafer, K., and Vafai, K., A critical synthesis of thermophysical characteristics of

nanofluids, International Journal of Heat and Mass Transfer, vol. 54, no. 19-20, pp. 4410–

4428, 2011.

[106] Yu, W., Xie, H., Li, Y., and Chen, L., Experimental investigation on thermal conductivity

and viscosity of aluminum nitride nanofluid, Particuology, vol. 9, no. 2, pp. 187–191, 2011.

[107] Halelfadl, S., Estellé, P., Aladag, B., Doner, N., and Maré, T., Viscosity of carbon nanotubes

water-based nanofluids: Influence of concentration and temperature, International Journal

of Thermal Sciences, vol. 71, pp. 111–117, 2013.

62
[108] Aladag, B., Halelfadl, S., Doner, N., Maré, T., Duret, S., Estellé, P., and Haleifadl, S.,

Experimental investigations of the viscosity of nanofluids at low temperatures, Applied

energy, vol. 97, pp. 876–880, 2012.

[109] Syam Sundar, L., Singh, M.K., and Sousa, A.C.M., Investigation of thermal conductivity

and viscosity of Fe3O4 nanofluid for heat transfer applications, International

Communications in Heat and Mass Transfer, vol. 44, pp. 7–14, 2013.

[110] Phuoc, T.S.X., and Massoudi, M., Experimental observations of the effects of shear rates

and particle concentration on the viscosity of Fe2O3-deionized water nanofluids, Internationl

Journal of Thermal Sciences, vol. 48, no. 7, pp. 1294–1301, 2009.

[111] Corcione, M., Empirical correlating equations for predicting the effective thtermal

conductivity and dynamic viscosity of nanofluids, Energy Conversion and Management,

vol. 52, no. 1, pp. 789–793, 2011.

[112] Sekhar, Y.R., and Sharma, K. V., Study of viscosity and specific heat capacity

characteristics of water-based Al2O3 nanofluids at low particle concentrations, Journal of

Experimental Nanoscience, no. june 2014, pp. 1–17, 2013.

[113] Kitano, T., Kataoka, T., and Shirota, T., An empirical equation of the relative viscosity of

polymer melts filled with various inorganic fillers, Rheol. Acta, vol. 20, pp. 207–209, 1981.

[114] Bobbo, S., Fedele, L., Benetti, A., Colla, L., Fabrizio, M., Pagura, C., and Barison, S.,

Viscosity of water based SWCNH and TiO2 nanofluids, Experimental Thermal and Fluid

Science, vol. 36, pp. 65–71, 2012.

[115] Paul, G., Chopkar, M., Manna, I., and Das, P.K., Techniques for measuring the thermal

conductivity of nanofluids: A review, Renewable and Sustainable Energy Reviews, vol. 14,

no. 7, pp. 1913–1924, 2010.

[116] Sundar, L.S., Farooky, M.H., Sarada, S.N., and Singh, M.K., Experimental thermal

conductivity of ethylene glycol and water mixture based low volume concentration of Al2O3

63
 and CuO nanofluids, International Communications in Heat and Mass Transfer, vol. 41, no.

0, pp. 41–46, 2013.

[117] Vajjha, R.S., and Das, D.K., Experimental determination of thermal conductivity of three

nanofluids and development of new correlations, International Journal of Heat and Mass

Transfer, vol. 52, pp. 4675–4682, 2009.

[118] Mansour, R.B., Galanis, N., and Nguyen, C.T., Experimental study of mixed convection

with water–Al2O3 nanofluid in inclined tube with uniform wall heat flux, International

Journal of Thermal Sciences, vol. 50, no. 3, pp. 403–410, 2011.

[119] Fotukian, S.M., and Esfahany, M.N., Experimental study of turbulent convective heat

transfer and pressure drop of dilute CuO/water nanofluid inside a circular tube, International

Communications in Heat and Mass Transfer, vol. 37, no. 2, pp. 214–219, 2010.

[120] Liu, Z.-H., and Li, Y.-Y., A new frontier of nanofluid research – Application of nanofluids

in heat pipes, International Journal of Heat and Mass Transfer, vol. 55, no. 23-24, pp.

6786–6797, 2012.

[121] Wang, P.-Y., Chen, X.-J., Liu, Z.-H., and Liu, Y.-P., Application of nanofluid in an inclined

mesh wicked heat pipes, Thermochimica Acta, vol. 539, pp. 100–108, 2012.

[122] Hajian, R., Layeghi, M., and Sani, K.A., Experimental study of nanofluid effects on the

thermal performance with response time of heat pipe, Energy Conversion and Management,

vol. 56, no. 0, pp. 63–68, 2012.

[123] Saidur, R., Meng, T.C., Said, Z., Hasanuzzaman, M., and Kamyar, A., Evaluation of the

effect of nanofluid-based absorbers on direct solar collector, International Journal of Heat

and Mass Transfer, vol. 55, pp. 5899–5907, 2012.

[124] Yousefi, T., Shojaeizadeh, E., Veysi, F., and Zinadini, S., An experimental investigation on

the effect of pH variation of MWCNT-H2O nanofluid on the efficiency of a flat-plate solar

collector, Solar Energy, vol. 86, no. 2, pp. 771–779, 2012.

64
[125] Yousefi, T., Veisy, F., Shojaeizadeh, E., Zinadini, S., and Veysi, F., An experimental

investigation on the effect of MWCNT-H2O nanofluid on the efficiency of flat-plate solar

collectors, Experimental Thermal and Fluid Science, vol. 39, no. 0, pp. 207–212, 2012.

[126] Yousefi, T., Veysi, F., Shojaeizadeh, E., and Zinadini, S., An experimental investigation on

the effect of Al2O3-H2O nanofluid on the efficiency of flat-plate solar collectors, Renewable

Energy, vol. 39, no. 1, pp. 293–298, 2012.

[127] Nieh, H.-M., Teng, T.-P., and Yu, C.-C., Enhanced heat dissipation of a radiator using oxide

nano-coolant, International Journal of Thermal Sciences, vol. 77, pp. 252–261, 2014.

[128] Murshed, S.M.S., Leong, K.C., and Yang, C., Determination of the effective thermal

diffusivity of nanofluids by the double hot-wire technique, Journal of Physics D: Applied

Physics, vol. 39, no. 24, pp. 5316–5322, 2006.

[129] White, S.B., Shih, A.J., and Pipe, K.P., Investigation of the electrical conductivity of

propylene glycol-based ZnO nanofluids, Nanoscale Research Letters, vol. 6, pp. 346–350,

2011.

[130] Rella, R., Spadavecchia, J., Menera, M.G., Capone, S., and Taurino, A., Acetone and

ethanol solid-state gas sensor based on TiO2 nanoparticles thin film deposited by matrix

assisted pulsed laser evaporation, Sensors and Actuators B, vol. 127, pp. 426–431, 2007.

[131] Madaria, A.R., Kumar, A., Ishikawa, F.N., and Zhou, C., Uniform, highly conductive, and

patterned transparent films of a percolating silver nanowire network on rigid and flexible

substrates using a dry transfer technique, Nano Research, vol. 3, no. 8, pp. 564–573, 2010.

[132] Anoop, K.B., Kabelac, S., Sundararajan, T., and Das, S.K., Rheological and flow

characteristics of nanofluids: Influence of electroviscous effects and particle agglomeration,

Journal of Applied Physics, vol. 106, no. 3, pp. 034909 1–7, 2009.

[133] Janot, R., and Guérard, D., One-step synthesis of maghemite nanometric powers by ball

milling, Journal of Alloys and Compounds, vol. 333, pp. 302–307, 2002.

65
[134] Chin, P.P., Ding, J., Yi, J.B., and Liu, B.H., Synthesis of FeS2 and FeS nanoparticles by

high-energy mechanical milling and mechanochemical processing, Journal of Alloys and

Compound, vol. 390, pp. 255–260, 2005.

[135] Lam, C., Zhang, Y.F., Tang, Y.H., Lee, C.S., Bello, I., and Lee, S.T., Large-scale synthesis

of ultrafine Si nanoparticles by ball milling, Journal of Crystal Growth, vol. 220, pp. 466–

470, 2000.

[136] Franzel, L., Bertino, M.F., Huba, Z.J., and Carpenter, E.E., Synthesis of magnetic

nanoparticles by paulsed laser ablation, Applied Surface Science, vol. 261, pp. 332–336,

2012.

[137] Cristoforetti, G., Pitzalis, E., Spiniello, R., Ishak, R., Giammanco, F., Muniz-Miranda, M.,

and Caprali, S., Physic-chemical properties of Pd nanoparticles produced by Pulsed Laser

Ablation in different organic solvents, Applied Surface Science, vol. 258, pp. 3289–3297,

2012.

[138] Mutisya, S., Franzel, L., Barnstein, B.O., Faber, T.W., Ryann, J.J., and Bertino, M.F.,

Comparison of in situ and ex situ bioconjugate of Au nanoparticles generated by laser

ablation, Applied Surface Science, vol. 264, pp. 27–30, 2013.

[139] Manera, M.G., Taurino, A., Catalano, M., Rella, R., Caricato, A.P., and Buonsanti, R.,

Enhancement of the optically activated NO2 gas sensing response of brookite Ti2

nanorods/nanoparticles thin films deposited by matrix-assisted pulsed-laser evaporation,

Sensors and Actuator B: Chemical, vol. 161, pp. 869–879, 2012.

[140] Singh, V., and Chauhan, P., Structural and optical characterization of CdS nanoparticles

prepared by chemical precipitation, Journal of Physics and Chemistry of Solids, vol. 70, pp.

1074–1079, 2009.

[141] Chen, J.-F., Wang, Y.-H., Guo, F., Wang, X.-M., and Zheng, C., Synthesis of Nanoparticles

with Novel Technology: High-Gravity Reactive Precipitation, Industrial & Engineering

Chemistry Research, vol. 39, no. 4, pp. 948–954, 2000.

66
[142] Hsu, W.-C., Chen, S.., Kuo, P.., Lie, C.., and Tsai, W.., Preparation of NiCuZn ferrite

nanoparticles from chemical co-precipitation method and the magnetic properties after

sintering, Materials Science and Engineering: B, vol. 111, no. 2-3, pp. 142–149, 2004.

[143] Zhu, J., Liu, S., Palchik, O., Koltypin, Y., and Gedanken, A., Shape-controlled synthesis of

silver nanoparticles by pulse sonoelectrochemical methods, Langmuir, no. 23, pp. 6396–

6399, 2000.

[144] Mastai, Y., Polsky, R., Koltypin, Y., Gedanken, A., and Hodes, G., Pulsed

sonoelectrochemical synthesis of cadmium selenide nanoparticles, J. Am. Chem. Soc., vol.

121, pp. 10047–10052, 1999.

[145] Zhu, J., Aruna, S.T., Koltypin, Y., and Gedanken, A., A Novel Method for the preparation of

Lead Selenide: Pulse Sonoelectrochemical Synthesis of Lead Selenide Nanoparticles, Chem.

Mater., vol. 12, pp. 143–147, 2000.

[146] Zin, V., Pollet, B.G., and Dabalà, M., Sonoelectrochemical (20kHz) production of platinum

nanoparticles from aqueous solutions, Electrochimica Acta, vol. 54, no. 28, pp. 7201–7206,

2009.

[147] Lei, H., Tang, Y.-J., Wei, J.-J., Li, J., Li, X.-B., and Shi, H.-L., Synthesis of tungsten

nanoparticles by sonoelectrochemistry., Ultrasonics sonochemistry, vol. 14, no. 1, pp. 81–3,

2007.

[148] Sáez, V., and Mason, T.J., Sonoelectrochemical Synthesis of Nanoparticles, molecules, vol.

14, pp. 4284–4299, 2009.

[149] Mueller, R., Mädler, L., and Pratsinis, S.E., Nanoparticle synthesis at high production rates

by flame pyrolysis, Chemical Engineering Science, vol. 58, pp. 1969–1976, 2003.

[150] Sahm, T., Mädler, L., Gurlo, A., Barsan, N., Pratsinis, S.E., and Weimar, U., Flame spray

synthesis of tin dioxide nanoparticles for gas sensing, Sensors and Actuators B, vol. 98, pp.

148–153, 2004.

67
[151] Kong, J., Cassell, A., and Dai, H., Chemical vapor deposition of methane for single-walled

carbon nanotubes, Chemical Physics Letters, no. august, pp. 567–574, 1998.

[152] Colomer, J.-F., Stephan, C., Lefrant, S., Tendeloo, G. V, Willems, I., Kónya, Z., Fonseca,

A., Laurent, C., and Nagy, J.B., Large-scale synthesis of single-wall carbon nanotubes by

catalytic chemical vapor deposition (CCVD) method, Chemical Physics Letters, vol. 317,

pp. 83–89, 2000.

[153] Nakaso, K., Han, B., Ahn, K.H., Choi, M., and Okuyama, K., Synthesis of non-

agglomerated nanoparticles by an electrospray assisted chemical vapour depostion (ES-

CVD) method, Aerosol Science, vol. 34, pp. 869–881, 2003.

[154] Sahoo, P.K., Kamal, S.S.K., Premkumar, M., Kumar, T.J., Sreedhar, B., Singh, A.K.,

Srivastava, S.K., and Sekhar, K.C., Synthesis of tungsten nanoparticles by solvothermal

decomposition of tungsten hexacarbonyl, Int. Journal of Refractory Metals and Hard

Materials, vol. 27, pp. 567–574, 2009.

[155] Rao, J.P., and Geckeler, K.E., Polymer nanoparticles: Preparation techniwues and size-

control parameters, Progress in Polymer Science, vol. 36, pp. 887–913, 2011.

[156] Koch, C.C., The synthesis and structure of nanocrystalline materials produced by

mechanical attrition: A review, NanoStructured Materials, vol. 2, pp. 109–129, 1993.

[157] Fedele, L., Colla, L., Bobbo, S., Barison, S., and Agresti, F., Experimental stability analysis

of different water-based nanofluids., Nanoscale research letters, vol. 6, no. 1, pp. 300, 2011.

[158] Zamzamian, A., Oskouie, S.N., Doosthoseini, A., Joneidi, A., and Pazouki, M.,

Experimental investigation of forced convective heat transfer coefficient in nanofluids of

Al2O3/EG and CuO/EG in a double pipe and plate heat exchangers under turbulent flow,

Experimental Thermal and Fluid Science, vol. 35, no. 3, pp. 495–502, 2011.

[159] Singh, A.K., and Raykar, V.S., Microwave synthesis of silver nanofluids with

polyvinylpyrrolidone (PVP) and their transport properties, Colloid and Polymer Science,

vol. 286, no. 14-15, pp. 1667–1673, 2008.

68
[160] Kim, D., Hwang, Y., Cheong, S.I., Lee, J.K., Hong, D., Moon, S., Lee, J.E., and Kim, S.H.,

Production and characterization of carbon nano colloid via one-step electrochemical method,

Journal of Nanoparticle Research, vol. 10, no. 7, pp. 1121–1128, 2008.

[161] Singaravelu, G., Arockiamary, J., Ganesh, K., and Govindaraju, K., A novel extracellular

synthesis of monodispersed gold nanoparticles using marine alga, Sargassum wightii grevile,

Colloids and Surface B: Biointerfaces, vol. 28, pp. 313–318, 2003.

[162] Ahmad, A., Mukherjee, P., Senapati, S., Mandal, D., Khan, M.I., Kumar, R., and Sastry, M.,

Extracellular biosynthesis of silver nanoparticles using the fungus Fusarium osysporum,

Colloids and Surfaces B: Biointerfaces, vol. 28, pp. 313–318, 2007.

[163] Bhainsa, K.C., and D\rqSouza, S.F., Extracellular biosynthesis of silver nanoparticles using

the fungus Aspergillus fumigatus, Colloids Surfaces B: Biointerfaces, vol. 47, pp. 160–164,

2006.

[164] Elumalai, E.K., Prasad, T.N.V.K. V, Hemachandran, J., and Therasa, S. V, Extracellular

synthesis of silver nanoparticles using leaves of Euphorbia hirta and their antibacterial

activities, Jorunal of Pharmaceutical Sciences and Research, vol. 2, no. 9, pp. 549–554,

2010.

[165] Krishnaraj, C., Jagan, E.G., Rajasekar, S., Selvakumar, P., Kalaichelvan, P.T., and Mohan,

N., Synthesis of silver nanoparticles using Acalypha indica leaf extracts and its antibacterial

activity against water borne pathogens, Colloids and Surface B: Biointerfaces, vol. 76, pp.

50–56, 2010.

[166] Huang, J., Li, Q., Sun, D., Lu, Y., Su, Y., Yang, X., Wang, H., Wang, Y., Shao, W., He, N.,

Hong, J., and Chen, C., Biosynthesis of silver and gold nanoparticles by novel sundried

Cinnamomum camphora leaf, Nanotechnology, vol. 18, no. 10, pp. 105104, 2007.

[167] Narayanan, K.B., and Sakthivel, N., Coriander leaf mediated biosynthesis of gold

nanoparticles, Materials Letters, vol. 62, no. 30, pp. 4588–4590, 2008.

69
[168] Song, J.Y., Jang, H., and Kim, B.S., Biological synthesis of gold nanoparticles using

Magnolia kobus and Diopyros kaki leaf extracts, Process Biochemistry, vol. 44, pp. 1133–

1138, 2009.

[169] Shivaji, S., Madhu, S., and Singh, S., Extracellular synthesis of antibacterial silver

nanoparticle using psychrophilic bacteria, Process Biochemistry, vol. 46, pp. 1800–1807,

2011.

[170] He, S., Guo, Z., Zhang, Y., Zhang, S., Wang, J., and Gu, N., Biosynthesis of gold

nanoparticles using the bacteria Rhodopseudomonas capsulata, Materials Letters, vol. 61,

pp. 3984–3987, 2007.

[171] Husseiny, M.I., El-Aziz, M.A., Badr, Y., and Mahmoud, M.A., Biosynthesis of gold

nanoparticles using Pseudomonas aeruginosa, Spectrochimica Acta Part A, vol. 67, pp.

1003–1006, 2007.

[172] Agnihotri, M., Joshi, S., Kumar, A.R., Zinjarde, S., and Kulkarni, S., Biosynthesis of gold

nanoparticles by the tropical marine yeast Yarrowia lipolytica NCIM 3589, Materials

Letters, vol. 63, pp. 1231–1234, 2009.

[173] Apte, M., Girme, G., Nair, R., Bankar, A., Kumar, A.R., and Zinjarde, S., Melanin mediated

synthesis of gold nanoparticles by Yarrowia lipolytica, Materials Letters, vol. 95, pp. 149–

152, 2013.

[174] Krumov, N., Oder, S., Perner-Nochta, I., Angelov, A., and Posten, C., Accumulation of CdS

nanoparticles by yeast in a fed-batch bioprocess, Journal of Biotechnology, vol. 132, pp.

481–486, 2007.

[175] Jha, A.K., and Prasad, K., A green low-cost biosynthesis of Sb2O3 nanoparticles,

Biochemical Engineering Journal, vol. 43, no. 3, pp. 303–306, 2009.

[176] Inbakandan, D., Sivaleela, G., Magesh Peter, D., Kiurbagaran, R., Venkatesan, R., and

Ajmal Khan, S., Marine sponge extract assisted biosynthesis of silver nanoparticles,

Materials Letters, vol. 87, no. null, pp. 66–68, 2012.

70
[177] Thakkar, K.N., Mhatre, S.S., and Parikh, R.Y., Biological synthesis of metallic

nanoparticles, Nanomedicine: Nanotechnology, Biology, and Medicine, vol. 6, pp. 257–262,

2010.

[178] Prasher, R., Phelan, P.E., and Bhattacharya, P., Effect of aggregation kinetics on the thermal

conductivity of nanoscale colloidal solutions (nanofluid)., Nano letters, vol. 6, no. 7, pp.

1529–34, 2006.

[179] Samal, S., Satpati, B., and Chaira, D., Production and dispersion stability of ultrafine Al–Cu

alloy powder in base fluid, Journal of Alloys and Compounds, vol. 504, pp. S389–S394,

2010.

[180] Wang, X., Li, X., and Yang, S., Influence of pH and SDBS on the stability and thermal

conductivity of nanofluids, Energy & Fuels, vol. 23, no. 16, pp. 2684–2689, 2009.

[181] Shima, P.D., Philip, J., and Raj, B., Magnetically controllable nanofluid with tunable

thermal conductivity and viscosity, Applied Physics Letters, vol. 95, no. 13, pp. 133112,

2009.

[182] Lee, S.W., Park, S.D., Kang, S., Bang, I.C., and Kim, J.H., Investigation of viscosity and

thermal conductivity of SiC nanofluids for heat transfer applications, International Journal

of Heat and Mass Transfer, vol. 54, no. 1-3, pp. 433–438, 2011.

[183] Phuoc, T.X., Massoudi, M., and Chen, R.-H., Viscosity and thermal conductivity of

nanofluids containing multi-walled carbon nanotubes stabilized by chitosan, International

Journal of Thermal Sciences, vol. 50, no. 1, pp. 12–18, 2011.

[184] Utomo, A.T., Poth, H., Robbins, P.T., and Pacek, A.W., Experimental and theoretical

studies of thermal conductivity, viscosity and heat transfer coefficient of titania and alumina

nanofluids, International Journal of Heat and Mass Transfer, vol. 55, no. 25-26, pp. 7772–

7781, 2012.

71
[185] Kumaresan, V., and Velraj, R., Experimental investigation of the thermo-physical properties

of water–ethylene glycol mixture based CNT nanofluids, Thermochimica Acta, vol. 545, pp.

180–186, 2012.

[186] Chen, H., Ding, Y., and Lapkin, A., Rheological behaviour of nanofluids containing tube /

rod-like nanoparticles, Powder Technology, vol. 194, pp. 132–141, 2009.

[187] Yang, J.-C., Li, F.-C., Zhou, W.-W., He, Y.-R., and Jiang, B.-C., Experimental investigation

on the thermal conductivity and shear viscosity of viscoelastic-fluid-based nanofluids,

International Journal of Heat and Mass Transfer, vol. 55, no. 11-12, pp. 3160–3166, 2012.

[188] Syam Sundar, L., Venkata Ramana, E., Singh, M.K., and De Sousa, a. C.M., Viscosity of

low volume concentrations of magnetic Fe3O4 nanoparticles dispersed in ethylene glycol

and water mixture, Chemical Physics Letters, vol. 554, pp. 236–242, 2012.

[189] Adio, S.A., Sharifpur, M., and Meyer, J.P., Investigation into Effective Viscosity and

Electrical Conductivity of γ-Al2O3-Glycerol Nanofluids in Einstein Concentration Regime,

UKHTC 2013. pp. 1–13 2013.

[190] Murshed, S.M.S., Leong, K.C., and Yang, C., Investigations of thermal conductivity and

viscosity of nanofluids, International Journal of Thermal Sciences, vol. 47, no. 5, pp. 560–

568, 2008.

[191] Prasher, R., Song, D., and Wang, J., Measurements of nanofluid viscosity and its implication

for thermal application, Applied Physics Letters, vol. 89, pp. 133108–1–3, 2006.

[192] Das, S.K., Putra, N., and Roetzel, W., Pool boiling characteristics of nano-fluids,

International Journal of Heat and Mass Transfer, vol. 46, no. 5, pp. 851–862, 2003.

[193] Wang, X., Xu, X., and Choi, S.U.., Thermal Conductivity of Nanoparticle - Fluid Mixture,

Journal of Thermophysics and Heat Transfer, vol. 13, no. 4, pp. 474–480, 1999.

[194] Hosseini, M., and Ghader, S., A model for temperature and particle volume fraction effect

on nanofluid viscosity, Journal of Molecular Liquids, vol. 153, pp. 139–145, 2010.

72
[195] Suresh, S., Venkitaraj, K.P., Selvakumar, P., and Chandrasekar, M., Synthesis of Al2O3-

Cu/water hybrid nanofluids using two step method and its thermo physical properties,

Colloids and Surfaces A: Physicochem. Eng. Aspects, vol. 388, pp. 41–48, 2011.

[196] Larson, R.G., The structure and rheology of complex fluids, Oxford University Press, Inc.,

New York, 1999.

[197] Chen, H., Witharana, S., Jin, Y., Kim, C., and Ding, Y., Predicting thermal conductivity of

liquid suspensions of nanoparticles (nanofluids) based on rheology, Particuology, vol. 7, no.

2, pp. 151–157, 2009.

[198] Moradpour, H., Chapoy, A., and Tohidi, B., Bimodal model for predicting the emulsion-

hydrate mixture viscosity in high water cut systems, Fuel, vol. 90, no. 11, pp. 3343–3351,

2011.

[199] Rensing, P.J., Liberatore, M.W., Sum, A.K., Koh, C. a., and Dendy Sloan, E., Viscosity and

yield stresses of ice slurries formed in water-in-oil emulsions, Journal of Non-Newtonian

Fluid Mechanics, vol. 166, no. 14-15, pp. 859–866, 2011.

[200] Singh, N., Chand, G., and Kanagaraj, S., Investigation of Thermal Conductivity and

Viscosity of Carbon Nanotubes –, Heat Transfer Engineering, vol. 33, no. 9, pp. 821–827,

2012.

[201] Mahbubul, I.M., Khaleduzzaman, S.S., Saidur, R., and Amalina, M. a., Rheological behavior

of Al2O3/R141b nanorefrigerant, International Journal of Heat and Mass Transfer, vol. 73,

pp. 118–123, 2014.

[202] Yang, Y., Oztekin, A., Neti, S., and Mohapatra, S., Particle agglomeration and properties of

nanofluids, Journal of Nanoparticle Research, vol. 14, no. 5, pp. 852, 2012.

[203] Keblinski, P., Prasher, R., and Eapen, J., Thermal conductance of nanofluids: is the

controversy over?, J. Nanopart Res., vol. 10, pp. 1089–1097, 2008.

[204] Tseng, W.J., and Lin, C.L., Effect of dispersant on rheological behaviour of BaTiO3

poweders in ethano-isopropanol mixtures, Mater. Chem. Phys., vol. 80, pp. 232–238, 2003.

73
[205] Nguyen, C.T., Desgranges, F., Roy, G., Galanis, N., Maré, T., Boucher, S., and Angue

Mintsa, H., Temperature and particle-size dependent viscosity data for water-based

nanofluids – Hysteresis phenomenon, International Journal of Heat and Fluid Flow, vol. 28,

no. 6, pp. 1492–1506, 2007.

[206] He, Y., Jin, Y., Chen, H., Ding, Y., Cang, D., and Lu, H., Heat transfer and flow behaviour

of aqueous suspensions of TiO2 nanoparticles (nanofluids) flowing upward through a

vertical pipe, International Journal of Heat and Mass Transfer, vol. 50, no. 11-12, pp.

2272–2281, 2007.

[207] Namburu, P., and Kulkarni, D., Experimental investigation of viscosity and specific heat of

silicon dioxide nanofluids, Micro & Nano Letters, vol. 2, no. 3, pp. 67–71, 2007.

[208] Pastoriza-Gallego, M.J., Casanova, C., Legido, J.L., and Piñeiro, M.M., CuO in water

nanofluid: Influence of particle size and polydispersity on volumetric behaviour and

viscosity, Fluid Phase Equilibria, vol. 300, no. 1-2, pp. 188–196, 2011.

[209] Anoop, K.B., Sundararajan, T., and Das, S.K., Effect of particle size on the convective heat

transfer in nanofluid in the developing region, International Journal of Heat and Mass

Transfer, vol. 52, no. 9-10, pp. 2189–2195, 2009.

[210] Timofeeva, E. V, Smith, D.S., Yu, W., France, D.M., Singh, D., and Routbort, J.L., Particle

size and interfacial effects on thermo-physical and heat transfer characteristics of water-

based alpha-SiC nanofluids., Nanotechnology, vol. 21, no. 21, pp. 215703, 2010.

[211] Jia-Fei, Z., and Zhong-Yang, L., Dependence of nanofluid viscosity on particle size and pH

value, Chinese Physics Lett., vol. 26, no. 6, pp. 10–13, 2009.

[212] Rubio-Hernández, F.J., Ayúcar-Rubio, M.F., Vlaázquez-Navarro, J.F., and Galindo-Rosales,

F.J., Intrinsic viscosity of SiO2, Al2O3and TiO2 aqueous suspensions, Journal of Colloid and

Interface Science, vol. 298, pp. 967–972, 2006.

74
[213] Boluk, Y., Lahiji, R., Zhao, L., and McDermott, M.T., Suspension viscosities and shape

parameters of cellulose nanocrystals (CNC), Colloids and Surfaces A: Physicochem. Eng.

Aspects, vol. 337, pp. 297–303, 2011.

[214] Wierenga, A.M., and Philipse, A.P., Low-shear viscosities of (semi-)dilute, aqueous

dispersions of charged boehmite rods: dynamic scaling of double layer effects, Langmuir,

vol. 13, pp. 4574–4582, 1997.

[215] Wierenga, A.M., and Philipse, A.P., Low-shear viscosities of dilute dispersions of colloidal

rodlike silica particles in cyclohexane, Journal of Colloid and Interface Science, vol. 180,

pp. 360–370, 1996.

[216] Hwang, Y., Park, H.S., Lee, J.K., and Jung, W.H., Thermal conductivity and lubrication

characteristics of nanofluids, Current Applied Physics, vol. 6, pp. e67–e71, 2006.

[217] Vékás, L., Bica, D., and Avdeev, M., Magnetic nanoparticles and concentrated magnetic

nanofluids: synthesis, properties and some applications, China Particuology, vol. 5, pp. 43–

49, 2007.

[218] Sharifpur, M., Meyer, J.P., and Africa, S., Opportunities in Nanofluid Composites, The 3rd

International Coference on Composites: Characterization, Fabrication and Application

(CCFA-3). pp. 9–10. , Tehran, Iran, 2012.

[219] Charles, S., The preparation of magnetic fluids, Ferrofluids, pp. 3–18, 2003.

[220] Shima, P., and Philip, J., Tuning of thermal conductivity and rheology of nanofluids using

an external stimulus, The Journal of Physical Chemistry C, pp. 20097–20104, 2011.

[221] Andablo-Reyes, E., Hidalgo-Álvarez, R., and Vicente, J., Controlling friction using

magnetic nanofluids, Soft Matter, vol. 7, pp. 880–883, 2011.

[222] S S Papell, Low viscosity magnetic fluid obtained by the colloidal suspension of magnetic

particles, 1965.

[223] Vekas, L., Magnetic nanofluids properties and some applications, Romanian Journal of

Physics, vol. 49, pp. 707–721, 2004.

75
[224] Vekas, L., Bica, D., and Marinica, O., Magnetic nanofluids stabilized with various chain

length surfactants, Romanian Reports in Physics, vol. 58, no. 3, pp. 257–267, 2006.

[225] Shima, P., Philip, J., and Raj, B., Synthesis of aqueous and nonaqueous iron oxide

nanofluids and study of temperature dependence on thermal conductivity and viscosity, The

Journal of Physical Chemistry C, pp. 18825–18833, 2010.

[226] Wonterghem, J., Mørup, S., Charles, S., Wells, S., and Villadsen, J., Formation of a metallic

glass by thermal decomposition of Fe(CO)5, Physical review Letters, vol. 55, no. 4, 1985.

[227] Jolly, M., Bender, J.W., and Carlson, D.J., Properties and applications of commercial

magnetorheological fluids, SPIE 5th Annual Int. Symposium on Smart Structures and

Matherials. pp. 1–14. , San Diego, CA, 1998.

[228] Rudyak, V.Y., and Krasnolutskii, S.L., Dependence of the viscosity of nanofluids on

nanoparticle size and material, Physics Letters A, vol. 378, no. 26-27, pp. 1845–1849, 2014.

[229] Huseyin, K., and Muhammet, K., Prediction of thermal conductivity of ethylene glycol -

water solutions by using artificial neural networks, Applied Energy, vol. 86, pp. 2244–2248,

2009.

[230] Hojjat, M., Etemad, S.G., Bagheri, R., and Thibault, J., Thermal conductivity of non-

Newtonian nanofluids: Experimental data and modeling using neural network, International

Journal of Heat and Mass Transfer, vol. 54, pp. 1017–1023, 2011.

[231] Papari, M.M., Yousefi, F., Moghadasi, J., Karimi, J., and Campo, A., Modeling thermal

conductivity augmentation of nanofluids using diffusion neural networks, International

Journal of Thermal Sciences, vol. 50, pp. 44–52, 2011.

[232] Mehrabi, M., Sharifpur, M., and Meyer, J.P., Application of the FCM-based neuro-fuzzy

inference system and genetic algorithm-polynomial neural network approaches to modelling

the thermal conductivity of alumina--water nanofluids, International Communications in

Heat and Mass Transfer, vol. 39, pp. 971–997, 2012.

76
[233] Yousefi, F., Karimi, H., and Papari, M.M., Modeling viscosity of nanofluids using

diffusional neural networks, Journal of Molecular Liquids, vol. 175, pp. 85–90, 2012.

[234] Bahiraei, M., Hosseinalipour, S., Zabihi, K., and Taheran, E., Using Neural Network for

Determination of Viscosity in Water-TiO2 Nanofluids, Advances in Mechanical

Engineering, vol. 742680, no. 1-10, 2012.

[235] Karimi, H., and Yousefi, F., Application of artificial neural network–genetic algorithm

(ANN–GA) to correlation of density in nanofluids, Fluid Phase Equilibria, vol. 336, pp. 79–

83, 2012.

[236] Mehrabi, M., Sharifpur, M., and Meyer, J.P., Viscosity of nanofluids based on an artificial

intelligence model, International Communications in Heat and Mass Transfer, vol. 43, pp.

6–11, 2013.

[237] Atashrouz, S., Pazuki, G., and Alimoradi, Y., Estimation of the viscosity of nine nanofluids

using a hybrid GMDH-type neural network system, Fluid Phase Equilibria, vol. 372, pp.

43–48, 2014.

[238] Wong, K. V., and De Leon, O., Applications of Nanofluids: Current and Future, Advances in

Mechanical Engineering, vol. 2010, pp. 1–11, 2010.

[239] Yu, W., and Xie, H., A Review on Nanofluids: Preparation, Stability Mechanisms, and

Applications, Journal of Nanomaterials, vol. 2012, pp. 1–17, 2012.

[240] Huminic, G., and Huminic, A., Application of nanofluids in heat exchangers: A review,

Renewable and Sustainable Energy Reviews, vol. 16, no. 8, pp. 5625–5638, 2012.

[241] Weinberger, H., The Physics of the Solar Pond, vol. 8, no. 2, pp. 45–56, 1963.
[242] Al-Nimr, M. a., and Al-Dafaie, A.M.A., Using nanofluids in enhancing the performance of a

novel two-layer solar pond, Energy, vol. 68, pp. 318–326, 2014.

[243] Konakanchi, H., Vajjha, R.S., Chukwu, G., and Das, D.K., Measurements of pH of Three

Nanofluids and Development of New Correlations, Heat Transfer Engineering, no. june, pp.

81–90, 2014.

77
 Table 1 Summarised list of the available classical models
Investigators Classical Models Remarks
Einstein [47] eff  o 1 2.5  Established on extremely dilute suspension of rigid solid spheres and
non-interacting medium. Volume fraction of  ≤ 0.02. From the model,
it is clear that viscosity is a linear function of volume fraction.
Taylor [50]   2  An extension of model [47], for liquid containing drops of another liquid

   ' 3 o 
 in suspension. The liquid drops have been assumed spherical and, for
eff  o 1  2.5   sphericity to be maintained, there must be high surface tension.
   ' o  Therefore, this model is only valid when the condition above is met.

  

Brinkman [51] eff  o 1   2.5 This is an extension of [47] and for a volume fraction,  ≤ 0.04.

Vand [52] eff  o 1  2.5  7.348 2  ...

Mooney [53]  2.5  Formulated on the premise of [47] and the model is limited to rigid
eff  o exp   spherical particles. This is a semi-empirical model as the interaction
 1    data,  (crowding factor) was left to be obtained by empirical means.
Accounts for suspensions containing a wide spectrum of continuous size
distribution, i.e. for monodispersed suspension of finite concentration,
The crowding factor  will be different for particulate suspension of
two different diameters, see [53].
Roscoe [54] eff  o 1  1.35 2.5 This model equation was developed for spheres of equal size and high
concentration. For spheres of very diverse sizes, the viscosity is to be
predicted with  nf  o 1   
2.5
, this equation is valid for all
concentration and as the vol. concentration is tending towards zero, it
reduces to model [30].

Batchelor [55]  eff   o 1  2.5  6.5 2  Effect of interactions between particles was considered in the
development of this model. Within the limits of a very low particle
volume concentration, this model approaches model [47].

78
 Table 1 Continued

Investigators Classical Models Remarks

  m
Krieger and 
 eff   o 1 
 
 Covers virtually the whole spectrum of nanoparticles. m is the maximum
Dougherty [56]
 m  concentration at which flow can occur, and its value for high shear rate is
0.605.  is the intrinsic viscosity with a typical value of 2.5.

 eff   o 1  2.5   2  f  3 
Lundgren [57]  25  This model was proposed considering Brownian motion of isotropic
 4  suspension of rigid spherical particles. The resulting bulk stress on the
particles was taken into account.
Within the limits of a very low particle volume concentration, this model
approaches model [47].

Graham [58]    This model approaches [47] and [62] as the lower and upper limit of solid
  1 
eff  o 1  2.5  4.5   


 
  hs d p  2   hs d p  1   hs d p 
2
  

volume fraction tends to zero and infinity respectively. Cell-based theory
was used where spheres were arranged in equidistance to a central sphere.
The diameters of the spheres were assumed uniform and zero inertial,
Brownian motion, Van der Waals, and electroviscous forces were
considered.
Saito [59]  eff   o 2.5  14.1 2 
Hatchek [60]  eff   o 1 4.5  It is applicable for up to 40% solid concentration.
Thomas and  eff   o 1  2.5  4.83 2  6.4 3 
Muthukumar [61]
Frankel and Acrivos  1
 Developed using asymptotic technique to describe the viscosity of
9   m  3 
[62] eff  o suspension within the concentrated limit where maximum volume fraction
8 1    13 
  m  
is obtainable. Assume uniform solid sphere to complement Einstein’s work
from dilute to concentrated regime.

79
 Table 2 New theoretical models

Investigators New Models Remarks

 V d2
nf  o  p B p
Masoumi et al. [2] 72C Developed based on Brownian motion, considering five parameters
 (volumetric fraction, temperature, particle diameter, nanoparticle density
 3 dp and base fluid physical properties). Calculated a correction factor to take
6
care of simplification assumptions. Tested the models with nanofluids with
single- and two-base fluids.

Hosseini et al. [33]  T 
nf  o .exp  m   
  T0 
 d p   This model was formulated using dimensionless groups considering the
   h       viscosity of the base fluid, hydrodynamic volume fraction of nanoparticles,
 1  r  
diameter of nanoparticles, thickness of capping layer of the nanoparticles
and temperature as  1 ,  2 ,  3 and  4 respectively.
The dimensionless group is defined as
nf d T
1  ,  2  h ,  3  , and  4  , m is a system property constant,
bf 1 r T0
 , ,  are empirical constants obtainable from the experimental data.

Chen et al. [68]  m Modified [56], considering the effect of agglomeration, therefore proposed
  
nf  o 1  a  a as the effective volume fraction of agglomerates. The power in the
 m 
model       was evaluated to be -1.5125, a   (aa / a)3D where D is the
m

fractal index and the duo of a a and a are the radii of the agglomerates and
primary particles.

80
 Table 3 Summary of available empirical models

Investigators Empirical Models Concentration Size Temperature Remarks

(%) (nm) (oC)
Heyhat et al.  5.989  0.1-2 40 20-60  Al2O3-water nanofluid.
[21] nf  o T  Exp    Correlation coefficient of 0.99.
 0.278     Valid for the temperature range investigated.
Chandreasek     
n
0.33-5 43 25  Al2O3-water nanofluid.
ar et al. [32] nf  o 1  A     Model was developed based on mean free path
  1    
 between nanoparticles.
 Aand n were taken as 5 200 and 2.8 respectively.
Namburu et Log nf   Ae  BT 1-6.12 29 -35-50  CuO-EG/water (60:40) nanofluid.
al. [34]  Newtonian nanofluid.
A  1.8375 2  29.643  165.56  T is the absolute temperature in Kelvin.
R2 = 0.9873  A and B are empirical curve fit parameters and in this
B  4 10   110   1.86 10
6 2 3 2
case are functions of , with R2 = 0.99.
R2 = 0.988
Abareshi et  B 
 
0.125-0.75 25-50 30-70  α – Fe2O3-glycerol nanofluid.
al. [43] nf T   o T  Ae  T To   Non-Newtonian shear thinning nanofluid.
 Based on Vogel-Fulcher-Tammann (VTF) equation.
 nf was obtained at shear rate of 40 s-1
 A and To are fitting parameters of the shear viscosity.
 B is related to the free activation energy of the fluid
(empirically obtainable).
1.25
Kole and    aa  
1.3 0.5-2.5 40 5-80  CuO-gear oil nanofluid.
nf  o 1   Newtonian nanofluid ( = 0.5%).
Dey [44]  0.5  a   
   Non-Newtonian shear thinning (0.5≤  ≤ 2.5 %).
In  nf   A 
1000 B  T is the absolute temperature in Kelvin.
(T  C )  Aggregated size is 200-360 nm.
 A, B and C are empirical curve fit parameters with
deviation ~ 1.4%.
 nf was obtained at 30 oC.

81
 Table 3 Continued

Investigators Empirical Models Concentration Size Temperature Remarks

(%) (nm) (oC)
Chen et al.
[81]

nf  o 1  10.6  10.6 
2
 0-8a 25 20-60 

TiO2-EG nanofluid.
Newtonian nanofluid.
In  nf   A 
1000 B  Agglomerated size is 70-100 nm.
(T  C )  nf predicted the experimental data with R2 = 0.9989.
 A, B and C are empirical curve fit parameters with
deviation ~ 1.7%.
Tseng and nf  o  0.4513e0.6965 3-10 300 25  Ni-terpineol nanofluid.
Chen [82]  Dispersant concentration 0.5-10% of Ni weight.
 nf predicted the experimental with R2 = 0.9952.
Tseng and nf  o 13.47e35.98 5-12 7-20 25  TiO2-DI-water nanofluid.
Lin [83]  nf predicted the experimental with R2 = 0.98.
 nf was obtained at shear rate of 100 s-1.
Horri et al.    
2
 0-40  NiO/YSZ-furfuryl alcohol suspension.
[85] nf  o 1  2.5  A     Reconciles the models of Einstein [47], Chen et al. [68]
  m     and Browers’ model using mobility parameters  m   
 
.
 Shear rate ranging from 10-1000 s-1.
 A and m are obtainable from experimental data.
Sahoo et al. nf  Ae B / T C  1-10 53 -35-90  Al2O3-EG/water (60:40) nanofluid.
[86]  Non-Newtonian nanofluid shear thinning (-35 – 0oC).
 Newtonian nanofluid (0-90 oC).
 T is the absolute temperature in Kelvin.
 A, B and C are empirical curve fit parameters with R2 =
0.99.
Nguyen et al. nf  o  0.904e0.1483 0.15-13 47 22-25  Al2O3-water nanofluid.
[87]  No information on the dispersant used.
Nguyen et al. nf  o 1  0.025  0.015 2  0.15-12 36 22-25  Al2O3-water nanofluid.
[87]  No information on the dispersant used.

82
 Table 3 Continued

Investigators Empirical Models Concentration Size Temperature Remarks

(%) (nm) (oC)
Nguyen et al. 1.475  0.319  0.051 2  0.15-12 29 22-25  CuO-water nanofluid.
nf  o    No information on the dispersant used.
[87]  0.009
3

Nguyen et al. nf  o (1.125  0.0007T ) 1 and 4 29, 36 20-70  Al2O3-water and CuO-water nanofluid.
[87] nf  o (2.1275  0.0215T  0.0002T )
2
and 47
Garg et al. nf  o 1  11  0.5-2 200 –  Cu-EG nanofluid.
[90]  Newtonian nanofluid.
 A linear fit following Einstein’s model [47].
Godson et al. nf  o 1.005  0.497  0.1149 2  0.3-0.9 60 50-90  Ag-DI-water nanofluid.
[91]
Duangthongs nf  o  A  B  C 2  0.2-2 21 15, 25 and 35  TiO2-water nanofluid.
uk and  A, B and C are empirical constants obtained from curve
Wongwises fitting for the three different temperatures.
[92]
Chevalier et nf  o 1  8.3  0-6 190 nm 25  SiO2-Ethanol nanofluid.
al. [94]  Clearly a linear fit of the type of classical work of
Einstein [47].

2 
Chevalier et   (d )  SiO2-ethanol nanofluid.
al. [94] nf  o 1  a p  0-7 35 and 25  Correlated based on Krieger and Dougherty model.
 m  94  Fitted for the particle sizes with Da as the aggregated
 Da
1.2
 diameter (195 and 352 nm) corresponding to the
a (d p )     nanoparticle diameters.
 dp
   m is crowding factor as detailed by Prasher et al. [191].
Da is the aggregate diameter.
1 
In  nf   A    B
Kulkarni et 5-15 29 5-50 CuO-water nanofluid.
al. [101] T   Non-Newtonian nanofluid showing pseudoplastic and
shear-thinning behaviour.
 T is the absolute temperature in Kelvin.
 A and B are empirical constants and are dependent on the
particle volume fraction.

83
 Table 3 Continued

Investigators Empirical Models Concentration Size Temperature Remarks

(%) (nm) (oC)
Syam Sundar   
6.356 0-2% 13 20 – 60  Fe3O4-water nanofluid.
et al. [109] nf  o 1    Newtonian nanofluids.
 12.5 
  en  2   en  2  0-4
Phuoc and 1 1
20-40 25  Fe2O3-DI-water.
1
Massoudi nf          2  
2
 Modelled to express the dependent of viscosity on shear rate
      
 and volume fraction.
[110]
 Shear rates are 26.4, 79.2, 132, and 211 s-1.
 Non-Newtonian shear thinning at  ≥2 %.
 Dispersant.
Corcione  1   Correlated for a wide range of data from the literature.
nf  o  0.01-7.1 25-200 20-50
[111]  1  34.87(d p d f ) 0.3 1.03   Nanofluids consisting of Al2O3, TiO2, SiO2 and Cu
  nanoparticles were used.
1/3
 6M   d is the diameter.
d f  0.1 
 N f
   M is the molar mass of the base fluid.
 N is the Avogadro’s number.
0.05915
Sekhar and  T 
0.5602
 dp    
10.51
0.01-5 13-100 20-70  Al2O3-water nanofluid.
nf  0.9350 1  nf  1   1    Regression model based on experimental data from the
Sharma [112]  70   80   100 
literature.
 Deviations of -10% – +18%.
Kitano et al.  
2 0-6.2 – –  Modelled for polymer melts (for?) different inorganic fillers.
[113] nf  o 1    Based on Maron-Pierce’s equation r  1     ,  is a
2

 
constant for packing geometry.
  is related to packing geometry of various inorganic
materials that fill the polymer melts;   0.54  0.0125 p ,
where p is the average aspect ratio.
 The equation is only applicable above the shear) stress of 104
dyne/cm2.
Bobbo et al. nf  o 1  A  B 2  0.01-1% 21-60 10-80  TiO2-water nanofluids (21 nm).
[114]  SWCNT-water nanofluids (60 nm).
 Newtonian nanofluids.
 A and B are empirical parameters based on present
experimental data.
 Dispersant used (SDS and PEG).
84
Table 4 Overview of classification of methods of preparation of nanoparticles

Classification Methods Nanostructures Ref.

Physical Ball milling -Fe2O3 [133]
Fe2S4, Fe2S [134]
Si [135]
Pulsed laser ablation Fe2O4 and Fe3C [136]
Pd [137]
Au [138]
Laser deposition and MAPLE TiO [130, 139]
Chemical Chemical precipitation CdS [140]
CaCo3, Al(OH)3, [141]
SrCO3, NiCuZn [142]
Sonoelectrochemical synthesis Ag [143]
CdSe [144]
PbSe [145]
Pt [146]
Tungsten [147]
Spray pyrolysis SiO2 [149]
TiO [150]
Chemical vapour deposition SWCNT [151, 152]
SiO2 [153]
Thermal decomposition Tungsten [154]
Biological Algae Au [161]
Fungi Ag [162, 163]
Plant and plant extracts Ag [164–166]
Au [166–168]
Bacteria Ag [169]
Au [170, 171]
Yeast Au [172–175]
Marine sponge Ag [176]

85
 Table 5 Summary of equipment for nanoviscosity and their measurement bases

Nanofluids Equipment type & Manufacturer SP Temp. Basis of Ref.

(hr) (oC) measurement
Al2O3 – water Oscillation Viscometer VM – 10A 5 21-39 Resonating [17]
(CBC Co Ltd.) vibration
ZnO – EG LVDV – II + Viscometer 3 60-20 Shear rate [18]
(Brookfield Engr. Lab, USA)
Al2O3 – water LVDV – I – Prime C/P Viscometer 6 300a Shear rate [32]
(Brookfield Engr. Lab, USA)
ZnO – EG Glass Capillary Viscometer NA 25 oC Gravity [42]
and ZnO – G (Ostwald viscometer) induced
α–Fe2O3– LVDV – II + Pro EXTRA 0.5 NA Viscous drag [43]
glycol Viscometer
(Brookfield Engr. Lab, USA)
CuO–Gear LVDV – II – Pro Viscometer 4 -20-100 Viscous drag [44]
Oil (Brookfield Engr. Lab, USA)
EG – TiO2 Bohlin CVO Rheometer 20 20-60 Controlled [68]
(Malvern Instrument UK) shear stress
Al2O3 – LVDV – II – Pro Viscometer 4 20-100 Shear rate [76]
water/EG (Brookfield Engr. Lab, USA)
TiO2 – EG Bohlin CVO Rheometer 20 10-90 Controlled [81]
(Malvern Instrument UK) shear stress
Ni– tevpineol VT550 Viscometer 24 25 Shear rate [82]
(Gerbruder HAAKE Gmbh,
Germany)
FA – HAAKE Mars III Rheometer 24 22 ± 1 Shear rate [85]
NiO/YSZ (Thermo Fisher Sci. Inc.)
Al2O3 – LVDV – II + Viscometer 1.5 -35-90 Shear rate [86]
water/EG (Brookfield Engr. Lab, USA)
Al2O3 – water ViscoLab450 Model NA 20-85 Piston-type [87]
(Cambridge Applied Systems, (Couette flow )
USA)
SiO2 – Capillary microviscometer - - - [94]
ethanol
SiC – DI- SV-10 Vibro-viscometer 12 20 – 80 Resonating [182]
water (A&D Company, Japan) vibration
Fe2O3 – LDDV – II Pro Viscometer NA -20-150 Viscous drag [183]
DDW (Brookfield Engr. Lab, USA)
Al2O3 – water AR 1000 Rheometer 3min 20 ± 0.1 Shear rate [184]
TiO2 – water (TA Instrument, USA)
CNT – Bohlin CVO Rheometer NA 0 – 40 Controlled [185]
water/EG (Malvern Instrument, UK) shear stress
EG – TNT Bohlin CVO Rheometer NA 20 – 60 Controlled [186]
(Malvern Instrument, UK) shear stress
Cu - VEF Kinexus Pro 3 225-325* Controlled [187]
(Malvern Instrument, UK) shear stress
*
Temperature is given in oK, SP – Sonication period, EG – Ethylene Glycol, G – Glycerol, DDW
– Double Distilled Water, VEF – viscoelastic fluid.
86
 List of figure captions

Figure 1 Inconsistency in suspension viscosity predictions by different available models. Al2O3-

DI-water nanofluids prediction at 20 oC. Insets at points (1) and (2) depict the level of discordance

in the predicted relative viscosity values even for models built around particle volume

concentration.

Figure 2 Underprediction of Al2O3-DI-water nanofluids by classical models.

Figure 3 Underprediction of TiO2-DI-water nanofluids by classical models.

Figure 4 Instability sequence in nanofluids. At time to, the nanofluid is stable just after

preparation by ultrasonication or HPH, at t1, flocculation sets in and degenerates to agglomeration

at t2, which finally sediments at time t3. As  also increases, the tendency of the instability

sequence is high.

Figure 5 Effect of temperature on the viscosity of Al2O3-glycerol nanofluid.

Figure 6 Effect of shear rate on the rheology of suspension: (a) the stress (  )-shear rate (  ) curve

of TiO2-water nanofluid at different particle volume fractions; (b) the suspension viscosity (  nf )-

shear rate (  ) curve of TiO2-water nanofluid at different particle volume fractions [83].

Figure 7 Effect of shearing time water-based nanofluids of CNT and Al2O3 at 5 oC: (a) CNT –

water nanofluid showing shear thinning, thixotropic; (b) Al2O3-water nanofluid showing shear

thickening, thixotropic [108].

Figure 8. Effect of nanoparticle size on the relative viscosity of SiO2-DI water nanofluid [211].

Figure 9 Criteria and selection flow chart for the implementation of search and sort algorithm.

87
Figure 10 Generic algorithms for selection of appropriate nanofluid viscosity models. R, S,…V

are available nanofluid viscosity models with attributes R1, R2, …V5, V6 and sub-array attributes

r1, r2, …v5,v6.

88
Figure 1 Inconsistency in suspension viscosity predictions by different available models. Al2O3-

DI-water nanofluids prediction at 20 oC. Insets at points (1) and (2) depict the level of discordance

in the predicted relative viscosity values even for models built around particle volume

concentration.

89
Figure 2 Underprediction of Al2O3-DI-water nanofluids by classical models.

90
Figure 3 Underprediction of TiO2-DI-water nanofluids by classical models.

91
 % Volume Concentration ()
Stable nanofluids
Stability marker

Sedimented nanofluids with

supernantant

Time
to t1 t2 t3

Figure 4 Instability sequence in nanofluids. At time to, the nanofluid is stable just after

preparation by ultrasonication or HPH, at t1, flocculation sets in and degenerates to agglomeration

at t3, which finally sediments at time t3. As  also increases, the tendency of the instability

sequence is high.

92
Figure 5 Effect of temperature on the viscosity of Al2O3-glycerol nanofluid [189].

93
 (a)

(b)

Newtonian
bed

Figure 6 Effect of shear rate on the rheology of suspension: (a) the shear stress (  )-shear rate (  )

curve of TiO2-water nanofluid at different particle volume fractions; (b) the nanofluid viscosity (

nf )-shear rate (  ) curve of TiO2-water nanofluid at different particle volume fractions [83].

94
 (a)

(b
)

Figure 7 Effect of shearing time water-based nanofluids of CNT and Al2O3 at 5 oC: (a) CNT-

water nanofluid showing shear thinning, thixotropic; (b) Al2O3-water nanofluid showing shear

thickening, thixotropic [108].

95
Figure 8. Effect of nanoparticle size on the relative viscosity of SiO2-DI-water nanofluid [211].

96
 List relevant models

Define arrays and sub-arrays

Define criteria for selection

Implement search and sort routine

Select appropriate model

Select another
model

Model performance evaluation

No Yes
Does it meet the
Model okay
performance criteria ?

Figure 9 Criteria and selection flow chart for the implementation of search and sort algorithm.

97




Model Attribute Definition Criteria Definition Sort Operation for Model Model Adaptation to Physical
Preference Problem

R1 R2 R3 R4 R5 R6

r5 r6

S1 S2 S3 S4 S5 S6
R S V Physical Problem

s1 s2 s3 s4

V1 V2 V3 V4 V5 V6

v1 v2
Priority Indices 1 R1 Priority Indices 2 R1
R1 R1
Selection Criteria R2 R2
R3 R3
R1: Error Margin Model
R2: Imposed Constraint R4 R4
R3: Specific use of Model R5 R1 R1 R5 R1 R1
R4: Base Mechanism R6 R6
R5: Range of Embodied R7 R7
Parameters
R6: Ease of Use R1 R1 R1 R1 R1 R1 R1 R1
R7: Dispersion Energy
Components Criterion-based Attribute Criterion-based Attribute
Ranking 1 Ranking 2

Figure 10 Generic algorithm for selection of appropriate nanofluid viscosity models. R, S,…V are

available nanofluid viscosity models with attributes R1, R2,…V5, V6 and sub-array attributes r1,

r2,…v5, v6.

98
 Josua Meyer obtained his BEng (cum laude) in 1984, MEng (cum laude) in

1986, and his PhD in 1988, all in mechanical engineering from the University of

Pretoria and is registered as a professional engineer. After his military service

(1988-1989), he accepted a position as Associate Professor in the Department of

Mechanical Engineering at the Potchefstroom University in 1990. He was

Acting Head and Professor in Mechanical Engineering before accepting a position as Professor in

the Department of Mechanical and Manufacturing Engineering at the Rand Afrikaans University

in 1994. He was Chairman of Mechanical Engineering from 1999 until the end of June 2002, after

which he was appointed Professor and Head of the Department of Mechanical and Aeronautical

Engineering at the University of Pretoria from 1 July 2002. At present, he is the Chair of the

School of Engineering. He specialises in heat transfer, fluid mechanics and thermodynamic

aspects of heating, ventilation and air-conditioning. He is the author and co-author of more than

450 articles, conference papers and patents and has received various prestigious awards for his

research. He is also a fellow or member of various professional institutes and societies such as the

South African Institute for Mechanical Engineers, South African Institute for Refrigeration and

Air-Conditioning, American Society for Mechanical Engineers, American Society for Air-

Conditioning, Refrigeration and Air-Conditioning, and is regularly invited as a keynote speaker at

local and international conferences. He has also received various teaching and exceptional

achiever awards. He is an associate editor of Heat Transfer Engineering and Editor of the Journal

of Porous Media.

Saheed A. Adio is currently a PhD student in the Department of Mechanical and

Aeronautical Engineering, University of Pretoria, South Africa. He obtained his

BSc(Hons) degrees in mechanical engineering at the Obafemi Awolowo

University, Ile-Ife, Nigeria in 2005 with a second class upper mark and after a few

99
years of working, he obtained his MSc in 2010 with distinction from the same university. He is

currently working on the mathematical modelling and experimental investigation into the effective

viscosity of nanofluids under the supervision of Dr M. Sharifpur and Professor J.P. Meyer.

Mohsen Sharifpur is a senior lecturer and also responsible for the nanofluids

research laboratory in the Department of Mechanical and Aeronautical

Engineering at the University of Pretoria. He received his BSc in mechanical

engineering from Shiraz University, Iran. He completed his MSc in nuclear

engineering, and then received a full scholarship for PhD study in mechanical engineering

(thermal fluid) from Eastern Mediterranean University. He was the only postgraduate student who

received four out of four for the CGPA when he received his PhD. He is the author and co-author

of more than 40 articles and conference papers. His research interests include convective

multiphase flow, thermal fluid behaviour of nanofluids, porous media, CFD and waste heat to

work in thermal systems. He is also a reviewer for notable accredited journals.

Paul N. Nwosu obtained a PhD degree in mechanical engineering at the

University of Nigeria in 2011. He did a stint as a postdoctoral fellow at the

Mechanical and Aeronautical Engineering Department (Thermofluids

Research Group), University of Pretoria, South Africa, and Energy

Technology Department (Internal Combustion Engine Research Group), Aalto University,

Finland. His research interests are in renewable energy systems, nanofluids, mechatronics (with

embedded software development), finite element modelling and computer simulation.

100